0% found this document useful (0 votes)

524 views219 pages

The Indiana College Mathematics Competition (2001-2023) : Adam Coffman Justin Gash Rick Gillman John Rickert

The document discusses the Indiana College Mathematics Competition (ICMC), which has been held annually since 1966, promoting teamwork and problem-solving among undergraduate students. It highlights the competition's structure, which includes a mix of easy and challenging problems from various mathematical topics, and emphasizes the importance of collaboration and problem-solving skills. Additionally, it provides historical context about the competition's founders, Peter Edson and Paul Mielke, and their contributions to mathematics education in Indiana.

Uploaded by

Minh Tuấn

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

524 views219 pages

The Indiana College Mathematics Competition (2001-2023) : Adam Coffman Justin Gash Rick Gillman John Rickert

Uploaded by

Minh Tuấn

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 219

Problem Books in Mathematics

Adam Coffman
Justin Gash
Rick Gillman
John Rickert

The Indiana
College
Mathematics
Competition
(2001–2023)
Celebrating the Teamwork Spirit and
the Peter Edson Trophy
Problem Books in Mathematics

Series Editor
Peter Winkler, Department of Mathematics, Dartmouth College, Hanover, NH, USA
Books in this series are devoted exclusively to problems - challenging, difficult, but
accessible problems. They are intended to help at all levels - in college, in graduate
school, and in the profession. Arthur Engels “Problem-Solving Strategies” is good
for elementary students and Richard Guys “Unsolved Problems in Number Theory”
is the classical advanced prototype. The series also features a number of successful
titles that prepare students for problem-solving competitions.
Adam Coffman • Justin Gash • Rick Gillman •
John Rickert

The Indiana College

Mathematics Competition
(2001–2023)
Celebrating the Teamwork Spirit
and the Peter Edson Trophy
Adam Coffman Justin Gash
Mathematical Sciences Mathematics
Purdue University Fort Wayne Franklin College
Fort Wayne, IN, USA Franklin, IN, USA

Rick Gillman John Rickert

Mathematics and Statistics Mathematics
Valparaiso University Rose–Hulman Institute of Technology
Valparaiso, IN, USA Terre Haute, IN, USA

ISSN 0941-3502 ISSN 2197-8506 (electronic)

Problem Books in Mathematics
ISBN 978-3-031-62767-5 ISBN 978-3-031-62768-2 (eBook)
https://doi.org/10.1007/978-3-031-62768-2
Mathematics Subject Classification: 00A07, 01A05, 97U40

© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland
AG 2024
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse
of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

If disposing of this product, please recycle the paper.

Preface

Every springtime in Indiana since 1966, students have gathered from across the state
to participate in the Indiana College Mathematics Competition.
At this event, also called “the Friendly Competition,” everyone meets for a brief
orientation, and then splits up into teams of three students, each getting two hours
to work together to solve undergraduate-level math problems. Then everyone comes
back together for dinner and discussion. The three top-scoring teams are announced
the next day.
Because the ICMC is intended to be a student activity where all the participants
can enjoy some success in problem solving, each year’s problem set has both easier
questions and more challenging ones, in a format where answers are to be written
and explained, with the opportunity to get partial credit for a good start or the right
general idea. The main topics are from the standard undergraduate curriculum for
math majors:

Calculus Combinatorics/Discrete Math

Linear Algebra Probability
.
Number Theory Geometry
Abstract Algebra Real Analysis

Some problems are like textbook exercises or entrance exam questions for graduate
study, while some others are elementary logic puzzles not using any advanced
terminology or techniques. Some are original inventions by the test writers, and
others are well-known, classic contest problems. This competition doesn’t require
Olympiad-level training and the memorization of obscure formulas or inequalities,
just a good grasp of the knowledge from college math classes, problem-solving
skills, and teamwork!
Readers interested in a specific topic can use the Index (starting on page 219) to
find problems from a subject area (calculus, number theory, . . . ) or solutions using
certain methods (induction, telescoping sums, . . . ). Each year’s problem set appears
as it did in the competition—with no subject labels, indication of difficulty, or hints

v
vi Preface

The student participants in the 2015 ICMC, held at Taylor University. Photo by Professor Jeremy
Case, Taylor University [TU]

other than those occasionally included by the problem authors. The problems and
solutions chapters are connected to each other by hyperlinks (in the electronic book)
and page number references.

Fort Wayne, IN, USA Adam Coffman

Franklin, IN, USA Justin Gash
Valparaiso, IN, USA Rick Gillman
Terre Haute, IN, USA John Rickert
Historical Remarks and
Acknowledgments

On April 27, 1965, Peter Edson, a trustee of Wabash College, sent its dean of faculty
a memorandum that included a newspaper clipping about a unique mathematics
competition that was held among high schools in New Jersey. In this competition,
each school entered a team that worked as a team on a set mathematics examination.
Edson wondered if anything of that sort was done at the college level. The dean,
as deans do, handed the question off to one of his faculty members, Paul Mielke,
who answered that he knew of no such competition but that he would be willing to
suggest the idea to his Indiana colleagues. Response to his letter was immediate and
favorable. Paul and others took up the challenge of creating such an exam, which
has resulted in nearly 60 years of shared experiences for undergraduate mathematics
students in the state of Indiana.
In this introduction, we briefly share the life stories of these two men and of the
Indiana College Mathematics Competition which they inspired and brought to life.

Peter Edson

One might assume that Peter Edson was a well-known (at least at the time)
mathematician intimately involved with college education in the State of Indiana.
Surprisingly, while the second half of that sentence is accurate, the first is not. Peter
Edson was a newspaper journalist, with no particular training in the mathematical
sciences.
He was born in Hartford City, Indiana, in 1896 to parents who were teachers,
with degrees from Hanover College, located in Madison, IN (Appendix). While
he was still in high school, Peter worked as a correspondent for the Fort Wayne
News-Sentinel . He was a bright, determined young man. In 1913, his high school
newspaper wrote that “Pete was the pilot [of the Junior class]. He was the class
grand problem solver and semi-official idea generator.”
After graduating from high school and matriculating at Wabash College, he
worked for the Crawfordsville Journal and the Indianapolis Star . He also served in

vii
viii Historical Remarks and Acknowledgments

Peter Edson, Fort Wayne

High School basketball
uniform. Photo courtesy of
the Edson family [ME23 ]

several editorial roles, including as Editor-in-Chief for the Wabash College student
newspaper, The Bachelor . He was a young man who clearly seemed to know what
he wanted to do with his life!
However, Peter’s life took a detour in 1917. That year he left Wabash College to
enter Reserve Officer Training prior the United States’ entry into WWI. Peter served
as a Lieutenant in the infantry until he was discharged in 1919. After the war, Edson
returned to Wabash College and his service on The Bachelor . A member of Phi Beta
Kappa, he graduated in 1920, worked briefly as a roustabout in the Ohio oilfields,
and then began working as the Sunday Editor at the Fort Wayne News-Sentinel .
This was the beginning of a life-long career as a professional journalist. Peter
earned a Master of Arts degree in 1925 from Harvard University. Concurrently, he
joined the Boston Post . From there, he moved to the New Haven Register and then
on to the Pittsburgh Press . By 1927 he was the editor for Every Week Magazine .
In 1932 he was appointed as the Editor-in-Chief of the Newspaper Enterprise
Association (NEA), a syndication service.
In 1941, Edson was promoted by the NEA to be its lead Washington correspon-
dent, a post he held until his retirement in 1964. In the archives of his papers, held by
the Wisconsin Historical Society [WHS], there are many congratulatory messages
Historical Remarks and Acknowledgments ix

Peter Edson, Wabash fraternity photo (at right, in white shirt). Photo courtesy of the Edson family
[ME23 ]

Peter Edson’s desk in Washington. Photo courtesy of the Edson family [ME23 ]

on this promotion and praise for his work as Editor-in-Chief. As he stepped into this
new position, he wrote that his vision was to write “a very informal piece, going
after the color and humor and the human interest of Washington. There is a gold
mine of this copy which no one seems to have staked a claim on, and I’d like to
make that mine, mine.”
x Historical Remarks and Acknowledgments

An early example of Peter’s careful, and popular, writing style is a 1941 column
about the Lend-Lease program, just passed by Congress, titled “How much is $1.3
billion? Breakdown of Lease-Lend Limit Shows Big Sum.” In the column, Edson
carefully explained what could be purchased with this amount of money. He begins
by noting that $500,000,000 goes, off the top, to Winston Churchill’s request for
“war tools.” Churchill also asked for 50 destroyers, costing $8,000,000 each for
another $400,000,000, leaving the remaining $400,000,000 for planes and other
ships.
Edson reports that the British were losing 68,000 tons of shipping per week and
could only replace half of it themselves. At the time, cargo ships were averaging
$333 per ton to build, but a recent contract demonstrated that this could be brought
down to $166 per ton. At the lower price, the US could build the needed tonnage for
a total of approximately $295,000,000. Then, given a price of $125,000 per patrol
bomber, we can supply 850 planes to the British with the remaining $105,000,000.
Edson reminds readers that the British could swap eight destroyers for a battle
ship, or 200 planes for a cruiser. And finally, he wryly notes that if you did not like
these numbers, then there is always next year’s budget.
Similarly, a 1942 column on the problems of the Japanese-American internment
program foreshadowed a dark chapter in American history, and a 1945 column on
post-war immigration laid out the pros and cons to generous immigration policies
following the war before coming down decisively in favor of immigration.
His work, and work ethic, quickly made Peter a nationally recognized journalist.
He won several journalism awards including the 1945 Sigma Delta Chi Award
for Journalism and the 1952 Headliners Award for Reporting. In 1948 he was
awarded the Raymond Clapper Memorial Award for excellence in Washington
Reporting. This award was identified by the White House Correspondents Asso-
ciation (WHCA) as its highest award after Clapper’s death in 1944. The Award was
given by the WHCA to a journalist or team for distinguished Washington reporting.
Peter Edson’s most notable journalistic achievement was his writing that broke
the Nixon slush fund scandal which led to Nixon’s now famous “Checkers” speech.
By the mid-1950s, his column was being carried in over 800 newspapers and being
read by more than 13 million people weekly. In a report on the reach of the NEA,
an association vice-president wrote that “this should make Edson far and away the
most widely printed Washington columnist.”
Peter was a modest man who was a firm believer in teamwork and in sharing
credit appropriately. For example, The Raymond Clapper Award came with a cash
prize which Peter promptly re-distributed to his more junior colleagues.
But how did Peter Edson get connected to an Indiana mathematics competition?
Clearly, he was a Hoosier. As noted above, he had heard of a team mathematics
competition for high school students in New Jersey. Peter asked the dean of the
faculty if there were anything like this in Indiana. And the rest is history, as one
might say.
Peter’s daughters also shared that he had designed the Washington DC home,
both the original home and an addition, that they grew up in. They also told the
story of Edson’s ability to modify a self-closing gate they saw in Williamsburg, VA,
Historical Remarks and Acknowledgments xi

for their home garden, and his fascination with the organization and efficiency of
the Disney World they visited on a family vacation [ME22 , MA].
Edson was deeply attached to his home state of Indiana. He returned frequently
for visits and never lost his Hoosier values of modesty and hard work. He served
on the Wabash College Board of Trustees from 1952 to 1964. The college awarded
him the 1955 Alumni Award of Merit and, in 1971, an honorary Doctor of Humane
Letters.
Peter died in 1977.

Paul Mielke

Paul Mielke was born in Racine, Wisconsin, in 1920, the only son in a family with
five daughters. In contrast to Peter Edson, his mathematical aptitude showed at an
early age while he spent time working with his father, who was an expert tool and die
maker. Paul was also an expert carpenter, personally creating the trophies awarded
for the competition during its early years. In addition to his skills with tools, Paul
was a very talented draftsman, producing beautiful drawings using India ink with
the stainless steel quills in his drafting set.
Paul eventually matriculated to Wabash College, where he earned an A.B. in
mathematics in 1942. Service in the Marine Corps during WWII interrupted his
study at Brown University, where he eventually, in 1946, earned a Sc. M. in Applied
Mathematics. He returned to Indiana, teaching briefly at Wabash College before
earning his Ph.D. in Mathematics at Purdue University in 1951.
Although his thesis was in mathematical analysis [M1 ] and much of his work
was in that area, Paul had a love of the classical subjects of number theory and
geometry, presumably drawing on the craftsmanship of his training as a child.
One interesting intersection of Paul’s practical and mathematical interests was in
compass and straightedge constructions. The one that impressed his son the most
was that he could construct three-dimensional paper models of the five Platonic
solids using nothing more than a T-square, a compass, a piece of heavy drafting
paper, a pair of scissors, and some Elmer’s glue [WCNews].
Mielke returned to Wabash College after completing his degree at Purdue. He
continued to work there throughout his professional career. While advancing in
faculty rank, Paul also founded and served as head of the college’s computer lab
in the early 1960s and also served as chair of the Department of Mathematics from
1963 to 1978. The college awarded him an honorary degree a few years after his
retirement in 1985.
As many mathematicians do, Paul took leaves from Wabash to engage in related
professional work. The years between 1952 and 1957 saw him in Seattle working
for Boeing Airline Company as a structural dynamics engineer. The opportunity
was attractive for a number of reasons: it gave him the chance to be on the leading
edge of applying the new technology of computers to solve real world problems; the
Korean War, which he feared would lead to WWIII, had just started and this was a
xii Historical Remarks and Acknowledgments

This photograph of Paul Mielke appears here courtesy of the Robert T. Ramsay Jr. Archival Center,
Wabash College, Crawfordsville, IN [WCLib]

chance to contribute to the national defense. The Boeing team built one of the first
industrial computer labs and used the machines to solve differential equations.
The year 1969 found him on leave to work as the director of the Mathematical
Association of America’s Committee on the Undergraduate Program in Mathemat-
ics, the body which sets the standards for the undergraduate curriculum. He was
active in the Association throughout his career, holding many leadership roles in the
Indiana Section of the MAA and being recognized for this work by receiving the
section’s first Distinguished Service Award. Besides leading the CUPM, Paul also
served on the Association’s Board of Directors and was an associate editor for the
American Mathematical Monthly from 1974 to 1978.
Paul published sparingly, but in quality publications, with articles in the Mathe-
matics Teacher [M2 ] and in the Two Year College Mathematics Journal (now the
College Mathematics Journal ) [M3 ]. However, his enthusiasm for teaching was
palpable and it is evident that many of his students felt the depth of his commitment
to teaching.
He was always prepared. For every textbook, he had completely worked out the
answer to every problem in every chapter. As Paul’s son relates, “He took special
satisfaction in the design of the new classrooms in Baxter with chalk boards on
two or three walls of the room and chairs that rotated in place. I recall several
lectures in which he had his presentation carefully planned so that he used all the
available chalkboard space and didn’t have to erase anything. He would complete the
peroration of his lecture and punctuate the conclusion with his chalk as he completed
the last proof at the bottom of the last available panel of the chalkboard.” [WCNews]
Historical Remarks and Acknowledgments xiii

This photograph of Paul

Mielke appears here courtesy
of the Robert T. Ramsay Jr.
Archival Center, Wabash
College, Crawfordsville, IN
[WCLib]

If Peter Edson was a journalist with hidden mathematical talents, Paul was a
mathematician with other hidden talents and interests. Besides the woodworking
skills mentioned above, his daughter wrote that her dad loved dancing and photog-
raphy, winning local prizes in both areas. More significantly, he was active in the
Civil Rights movement of the 1960s, founding and leading the Crawfordsville, IN,
branch of the NAACP as it raised funds and donations for the movement.
Paul passed away in 2008.

The Indiana College Mathematics Competition

The convergence of Peter Edson’s and Paul Mielke’s lives and interests is not quite
as surprising as it might seem. Both had life-long love of the state of Indiana,
and particularly of Wabash College which they attended and later served in some
capacity. Both men valued college education and the need to challenge young
men and women to excel. Further, they understood the value of teamwork and of
rewarding success.
The Indiana College Mathematics Competition (ICMC) was launched in 1965 by
the Indiana Section of the Mathematical Association under Mielke’s leadership. As
a small college competition (the competition was originally restricted to students
at the private colleges in Indiana), it was intended to serve as counterpoint to
the nationally distributed but notoriously difficult Putnam Exam. To this point,
the exam focused on mathematical topics in the undergraduate curriculum and
xiv Historical Remarks and Acknowledgments

This photograph of Paul Mielke appears here courtesy of the Robert T. Ramsay Jr. Archival Center,
Wabash College, Crawfordsville, IN [WCLib]

was intentionally designed to be completed by a team of students, rather than by

individuals. It retains these characteristics today. A brief history of the early years
of the ICMC written by Professor Mielke himself appears in [AFMC].
Typically, the ICMC is held on the first afternoon of the spring meeting of
the Indiana Section of the Mathematical Association of America (MAA). Student
teams check-in during the mid-afternoon after driving to the meeting site. After
general instructions are given, the teams are dispersed to their assigned work spaces,
often with students from the host institution showing them around campus. This
workspace is typically a classroom, but as the number of teams increases they are
often assigned to conference rooms—any place with a blackboard or a whiteboard!
They then have two hours to complete the exam and return it, all in time to attend
the conference banquet and listen to the keynote speaker.
A team of faculty who have volunteered to be graders work overnight to mark the
exams on a scale of .0−10 for each problem, with partial credit awarded according
to the solution guides from the problem writers and the impressions of the graders.
Winners are announced at the following day’s business meeting, held after morning
sessions which often include a workshop specifically designed for undergraduate
students.
The winning team is awarded the Peter Edson Trophy. Before his death, Paul
Mielke would handcraft this trophy from native Indiana wood.
Historical Remarks and Acknowledgments xv

The 2003 trophy for the first

place team from
Indiana–Purdue Fort Wayne

The ICMC’s strict two-hour limit, including walking time, puts an interesting
limit on the complexity and obscurity of the problems. Each individual exam reflects
the idiosyncrasies of the writers, and the difficulty of the problem sets can vary
significantly from year to year. In preparing for and participating in competitions
like the ICMC, students often learn about mathematics not covered in their college
courses. It seems unlikely that a single student would have taken courses in all of
the topics covered by the exam, implying that the “team” element of the competition
is essential. However, when viewed over the long-term history of the exam, the
problems reflect an evolving core of mathematical knowledge in the undergraduate
curriculum.
In the spring of 2020, the Indiana Section meeting, including the ICMC, was
abruptly but necessarily canceled, under the circumstances of the onset of the
COVID-19 pandemic. The 2021 ICMC event was temporarily re-organized into
a simultaneous, but remote, locally proctored format. In 2022, the competition
returned to its in-person format.

Acknowledging the Volunteers Who Organize the ICMC

The Mathematical Association of America, MAA, is a national professional orga-

nization founded in 1915 for college mathematics teachers and students, with over
25,000 faculty and student members, and a large budget and professional full-time
staff. The operations of each of its 29 regional Sections, including the Indiana
Section which has been active since 1924, are organized by volunteer members
with a small budget. The Indiana College Mathematics Competition has been the
main student activity at meetings of the Indiana Section since 1966; the competition
was quickly nicknamed the “Friendly” Competition because of its focus on solving
xvi Historical Remarks and Acknowledgments

mathematical problems with teamwork, and bringing faculty and students together
from around the state, rather than a focus on Olympiad-level individual competition.
We (the editors of this volume: AC, JG, RG, JR) would like to recognize the work
of many individuals who have contributed to the success of the Indiana College
Mathematics Competition over the past couple of decades.
These individuals include the mathematics professors who have served as the
Indiana Section’s Student Activities Coordinator each year, leading the organizing
of the ICMC event: Aaron Klebanoff (2000–2001), Mohammad Azarian (2001–
2007), Robert Talbert (2007–2010), Justin Gash (2010–2016), Paul Fonstad (2016–
2019), and Colin McKinney (2019–2023). The Coordinator is an officer on the
Executive Board of the Indiana Section of the MAA, and all the Board members
contribute to organizing the Section meeting at a different host institution every
year, from among the colleges and universities in Indiana (listed in Appendix).
Each year, the exam and a solution set are assembled by a committee from
the local host institution, or by an individual or team recruited by the Student
Activities Coordinator. The writers compile a list of problems, guided only by their
motivation to provide a fun and interesting student activity. In particular, the contest
writers have not been obliged to invent new, original problems, but have sometimes
copied or adapted problems from past ICMC events or similar contests, or drawn
from the math problem folklore. We would like to thank all these writers for their
creative contributions to the success of mathematics education and the mathematics
community in Indiana.
The Student Activities Coordinator also recruits graders, usually a group of
faculty attending the Spring Meeting, with each grader evaluating the teams’
answers to one or two problems and reporting their scores the morning after
the competition. The host institutions and local organizers of the spring Section
meetings all deserve acknowledgment. They provided the competitors with efficient
registration procedures, excellent campus guides, and many classrooms so that each
team could work in their own space. Without their help, this competition would not
run as smoothly as it has for the past 20 years.
And, finally, we would like to thank the many dedicated coaches and hundreds of
undergraduate students who have participated in the competition. They are the folks
who put the “friendly” in the name of the competition and bring joy to their home
institutions with their interest and enthusiasm in mathematical problem solving.

About This Volume

The problems and solutions from the first 35 years (1966–2000) of the competition
were gathered in the book A Friendly Mathematics Competition [AFMC]. Chap-
ter 47 of this work contains some corrections to that volume as well as new solutions
to some of the older problems. The problems in this collection that are repetitions
or variations of these older problems from [AFMC] are indicated in remarks in
the solution sets. Some of the problems here adapted from other sources are also
Historical Remarks and Acknowledgments xvii

marked with a link to the list of references, which gives the reader a sample of
problem books at a level similar to the ICMC [GKL, HWi, K1 , K2 , RL].
In this compilation (2001–2023), all the problems are copied as presented in the
contests, although a few statements have been edited for clarity or correctness, in
the interest of prioritizing usefulness to the reader over historical preservation. As
the solution sets provided by the exam authors at the time of the contests were
intended only as guides for the graders, some of those solutions were sketches
for experts rather than detailed, model solutions for students. So, many of the
solutions in Chaps. 24–45 are expanded versions of the originals distributed after
each competition. Some expanded, re-written, or alternate solutions are based on
the real-time answers from student contestants, or on feedback from reviewers well
after the contest, and some solutions are entirely new. For some problems, writing a
solution with the amount of detail appearing here may not be practical or strategic
for students in a two-hour contest situation. However, the ICMC exams are designed
to allow for partial credit to be assigned by the grader on a 10-point scale for each
problem, so a submitted solution judged to have some correct steps, a good partial
solution, or the right general idea can get some points.
The editors are pleased to thank the past and present members of the Executive
Board of the Indiana MAA for their support. In particular, Professor Vesna Kilibarda
of Indiana University Northwest was Indiana Section Chair in 2010–2011, and she
was an original member of the committee exploring the possibility of a sequel to
[AFMC].
We thank the competition’s most prolific contributors, Mike Axtell and Joe
Stickles, for sharing their original electronic files, which filled some gaps in our
records. The editors also thank Professor Steve Kennedy for his support during
early stages of this project, and Professor George Gilbert and several anonymous
reviewers, whose generous suggestions improved the content and contributed some
of the alternate solutions.
Credit for some photos appearing here is given in the appendix list of
references—photos appearing without credit were shared here by A. Coffman.
The historical sections on Edson and Mielke were prepared by R. Gillman, in
correspondence with members of the Edson family [ME22 , ME23 , MA], and with
the assistance of Elizabeth Swift, the archivist at Wabash College [WCLib].
Any author royalties or payments from sales of this volume will be donated to
the Indiana Section of the MAA to support future competitions and other student-
focused activities.
Contents

Part I Problems
1 2001 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2 2002 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3 2003 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
4 2004 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
5 2005 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
6 2006 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
7 2007 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
8 2008 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
9 2009 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
10 2010 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
11 2011 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
12 2012 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
13 2013 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
14 2014 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
15 2015 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
16 2016 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
17 2017 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
18 2018 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
19 2019 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
20 2020 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

xix
xx Contents

21 2021 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
22 2022 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
23 2023 Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

Part II Solutions
24 2001 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
25 2002 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
26 2003 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
27 2004 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
28 2005 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
29 2006 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
30 2007 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
31 2008 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
32 2009 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
33 2010 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
34 2011 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
35 2012 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
36 2013 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
37 2014 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
38 2015 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
39 2016 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
40 2017 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
41 2018 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
42 2019 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
43 2021 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
44 2022 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
45 2023 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193

Part III More History of the ICMC

46 Top Scoring Teams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
Contents xxi

47 Updates to A Friendly Mathematics Competition . . . . . . . . . . . . . . . . . . . . . . . . 207

Some Errata for the 1966–2000 Problem Book . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
New Solutions for Old Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208

Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
Location Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
References and Photo Credits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
Part I
Problems
Chapter 1
2001 Problems

P2001-1 Suppose that n is a positive integer, not a power of 2. Prove that n is the
sum of two or more consecutive positive integers.

P2001-2 Let .f (x) = x 1/x , .x > 0. Prove that f has exactly two inflection points.
x 1/x ⎛ ⎞
Hint: .f '' (x) = 4 (ln(x))2 + (2x − 2) ln(x) + (1 − 3x) .
x

P2001-3 A machine is required to drop k indistinguishable balls into n numbered

slots, .k ≥ n, and must drop at least one ball into each slot. How many outcomes
may arise?

P2001-4 There is a cubic polynomial f with integer coefficients and the property
that .f (cos(θ )) = cos(3θ ):

(a) Find f .
π
(b) Prove that .cos is irrational.
9

P2001-5 A circle .C1 of radius 1 is inside of and tangent to a circle .C2 of radius
2. .C1 rolls around the interior of .C2 without slipping, making continuous tangential
contact with .C2 , and returns to its initial position. Describe the locus of points traced
out by a distinguished point P on the circumference of .C1 .

⎲∞
| sin(n)|
P2001-6 Is . finite or infinite?
n
n=1
The 2001 Problem Solutions begin on page 65 in Chap. 24.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 3

A. Coffman et al., The Indiana College Mathematics Competition (2001–2023),
Problem Books in Mathematics, https://doi.org/10.1007/978-3-031-62768-2_1
Chapter 2
2002 Problems

P2002-1 Consider the expression

f (x)
. .
f (x) + f (b − x)

Note that depending on f , it is possible for the expression to be undefined for some
or all x in the interval .[0, b], so take care with the domains of the expressions
appearing in the following parts (a) and (b):

(a) Find a function .f (x) which is continuous on .[0, b] and not identically zero,
such that .f (x) + f (b − x) = 0 for all x in the interval .[0, b].
(b) For f continuous and positive on .[0, b], evaluate
⎰ b f (x)
. dx.
0 f (x) + f (b − x)

∑
P2002-2 Let . an be a series of strictly positive terms, and let

a1 + a2 + · · · + an
.bn = .
n
∑
Show . bn diverges.

P2002-3 Show that the sum of two consecutive odd primes has at least three (not
necessarily distinct) prime factors.

P2002-4 A university bookstore sold at least one mathematics textbook each day
for 100 consecutive days. During this time, the bookstore sold 140 mathematics

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 5

textbooks. Was there a period of consecutive days when exactly 59 mathematics

textbooks were sold?

P2002-5 Let .ϕ be a function from a set S with a binary operation (denoted by

juxtaposition) to the set of nonnegative integers such that .ϕ (xy) = ϕ (x) ϕ (y) for
all x and y in S:

(a) Supposing S is a group and the binary operation is the group operation, show
that either .ϕ (x) = 0 for all x in S, or .ϕ (x) = 1 for all x in S.
(b) Supposing .ϕ is a nonconstant function, S is a ring, and the binary operation is
the ring multiplication, show that .ϕ (0S ) = 0.

P2002-6 Show that a square inscribed in a triangle can enclose at most one half the
area of the triangle.
The 2002 Problem Solutions begin on page 71 in Chap. 25.
Chapter 3
2003 Problems

P2003-1 Let S be a set with a binary operation .∗ that has the property .(a ∗ b) ∗ (c ∗
d) = a ∗ d for all .a, b, c, d ∈ S. Prove these statements:

(a) If .a ∗ b = c, then .c ∗ c = c.
(b) If .a ∗ b = c, then .a ∗ y = c ∗ y for all .y ∈ S.

P2003-2 A tennis club invites .2k (.k ≥ 1) players of equal ability to compete in a
single-elimination tournament. Suppose that the probability of a competitor winning
his or her match is . 12 , and that the competitors are paired together in a random order
in the first round. Show that the probability that a particular pair of players will
1
compete against each other in this tournament is . k−1 .
2
P2003-3 Five points divide the unit circle into five equal arcs. Express ab in terms
of .sin 18◦ , where a and b are the lengths of the chords indicated in the picture here:

P2003-4 A .3×3 array of integers has the property that the products of the elements
in each of the three rows, in each of the three columns, and in each of the two

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 7

diagonals are all the same; call this product k. (We call such an array a multiplicative
magic square.) Prove that k is a perfect cube, and determine the element in the array
of which k is the cube.

P2003-5 Does the series

1 1 1 1 1 1 1 1
.1+ − + + − + + − ···
2 3 4 5 6 7 8 9
converge or diverge? Justify.

P2003-6 The Fibonacci sequence .f1 , f2 , f3 , . . . is defined by .f1 = f2 = 1,

fn = fn−1 + fn−2
.

for .n = 3, 4, 5, . . .. Thus, the sequence begins

.1, 1, 2, 3, 5, 8, 13, 21, 34, . . . .

Let
⎾ ⏋
11
.Q= :
10

(a) Prove
⎾ ⏋
fn+1 fn
.Q =
n
fn fn−1

for .n = 2, 3, 4, . . ..
(b) Establish the identity:

f3n = fn+1
.
3
+ fn3 − fn−1
3

for .n = 2, 3, 4, . . ..

The 2003 Problem Solutions begin on page 75 in Chap. 26.

Chapter 4
2004 Problems

P2004-1 Partition the set .{1, 2, 3, 4, 5} into two arbitrarily chosen sets. Prove that
one of the sets contains two numbers and their difference.

P2004-2 Suppose .a > 1:

⎲∞
2n
(a) Show the series . converges.
+1
n
a2
n=0
(b) Determine to what value this series converges.

(Note that a correct solution for part (b) necessarily solves part (a) as well.)

P2004-3 Let A be a .4 × 4 matrix such that each entry of A is either 2 or .−1. Let
d = det(A); clearly d is an integer. Show that d is divisible by 27.
.

P2004-4 Let .a1 , a2 , . . . , an be a finite sequence of real numbers. Form a sequence

of length .n − 1 by averaging two consecutive terms of the sequence:

a1 + a2 a2 + a3 an−2 + an−1 an−1 + an

. , ,..., , .
2 2 2 2
Continue this process of averaging two consecutive terms until you have only one
term left. Show that this final term is
n−1 ⎛
⎲ ⎞
n−1
ai+1
i
i=0
. .
2n−1

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 9

P2004-5 Let P be the center of a square with side .AC. Let B be a point in the
exterior of the square such that .ΔABC is a right triangle with hypotenuse .AC.
Prove: .BP bisects ./ ABC.

P2004-6 Two ferryboats start at the same instant from opposite sides of a river,
traveling across the water on routes at right angles to the parallel shores. Each travels
at a constant speed, but one is faster than the other. They pass at a point 720 yards
from the nearest shore. Both boats remain at their slips 10 min before starting back.
On their return trips, they meet 400 yards from the nearest shore. How wide is the
river?
The 2004 Problem Solutions begin on page 81 in Chap. 27.
Chapter 5
2005 Problems

P2005-1 Evaluate the integral

⎰ 1 | sin(n cos−1 (x))|
. √ dx.
−1 1 − x2

P2005-2 All points in the plane are colored in red, white, or blue. Prove that there
is at least one pair of points of the same color with distance between them one unit.

P2005-3 Find the limit of the sequence defined by

1 ⎲
n
an =
. ln(1 + kn).
n3
k=1

P2005-4 Let .Q+ be the set of all positive rational numbers and let “.∗” be an
operation on .Q+ that satisfies the following conditions for all .a, b, c, d ∈ Q+ :

(a ∗ b)(c ∗ d) = (ac) ∗ (bd),

a ∗ a = 1,
a ∗ 1 = a.

Compute the value of the expression .((6/5) ∗ (8/15)) ∗ 2.

P2005-5 Consider matrices A and B in .Mn (R) such that .A3 = A2 and .A+B = In .
Show that the matrix .AB + In is non-singular and find its inverse.

P2005-6 Determine the real constants .a, b, c, and p, such that

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 11

⎾√ ⏋ 7
. lim 9x 4 − 24x 3 + 6x 2 + 5 − (ax p + bx + c) = .
x→∞ 3

P2005-7 Let M and N be the midpoints of BC and CD in the parallelogram

AP
ABCD, and let P be the intersection of AM and BN. Determine the ratios .
AM
BP
and . .
BN

P2005-8 Given the integers x, y, and z, prove that if 25 divides the sum .x 5 +y 5 +z5 ,
then 25 divides at least one of the numbers .x 5 + y 5 , .x 5 + z5 , or .y 5 + z5 .
The 2005 Problem Solutions begin on page 85 in Chap. 28.
Chapter 6
2006 Problems

P2006-1 Find the maximum value of the function

√ √
.f (x) = 4x − x 2 + 12 − 10x − x 2 − 24.

P2006-2 Evaluate
∞
⎲ ⎛ ⎞
2
. arctan .
n2
n=1

Hint: It may be helpful to remember that

⎛ ⎞
α−β
. arctan(α) − arctan(β) = arctan .
1 + αβ

P2006-3 Let X be a continuous random variable having the probability density

function
⌠
1
, x≥1
.f (x) = x2
0, otherwise.
⎛⎾ √ ⏋ |⎾ √ ⏋ ⎞
|
Find .P 10X = 20 | 3 X = 3 (where .[x] denotes the greatest integer less
than or equal to x).

P2006-4 A bicyclist rides 18 miles in exactly 72 min. Assuming the bicyclist’s

position is a continuous, increasing function of time, prove that there exists a
contiguous 3-mile segment within this 18 miles that the rider completed in exactly
12 min.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 13

P2006-5 Let S be a set of real numbers that is closed under multiplication. Let T
and U be disjoint subsets of S whose union is S. Show that if the product of any
three (not necessarily distinct) elements of T is in T and the product of any three
elements of U is in U , then at least one of the two subsets T and U is closed under
multiplication.

P2006-6 Let r and s be nonzero integers. Prove that the equation

(r 2 − s 2 )x 2 − 4rsxy − (r 2 − s 2 )y 2 = 1
.

has no solutions in integers x and y.

P2006-7 Equilateral triangle ABC, with points P on side .AB and Q on side .AC,
has been creased and folded along .P Q so that vertex A lies at the point .A' on .BC.
Assuming .BA' = 1 and .A' C = 2, find the length of .P Q.

P2006-8 An evil genie pops out of a lamp and presents you with twelve identical-
looking stones. The genie tells you that one stone has an imperfection not visible to
the naked eye that causes the stone to be either slightly heavier or slightly lighter
than the others. The genie also presents you with a balance scale and says you may
use this scale at most three times. The genie will give you your heart’s desire if
you can correctly identify the imperfect stone and determine whether it is heavier or
lighter than the others.
The 2006 Problem Solutions begin on page 91 in Chap. 29.
Chapter 7
2007 Problems

P2007-1 Let p and q be distinct primes. Find a nonconstant polynomial with

√ √
integer coefficients that has . p + q as a root.

P2007-2 What is the value of the positive integer n for which the least common
multiple of 36 and n is 500 greater than the greatest common divisor of 36 and n?
⎰ 3x dt
P2007-3 Evaluate: . lim (x + 2) · √ .
x→∞ x t t4 + 1

P2007-4 Answer the following questions about prime numbers:

(a) Let p be a fixed prime and N a positive integer. Suppose an integer a is selected
at random from the set .{1, 2, 3, . . . , Np}. What is the probability that a is
divisible by p? (Think about the possible remainders when dividing by p.)
(b) For p, N as in (a), suppose two integers a and b are selected at random (inde-
pendently, with replacement) from .{1, 2, 3, . . . , Np}. What is the probability
that a and b are both divisible by p?
(c) Let .{2, 3, 5, 7, . . . , pk } be the list of the first k prime numbers, and denote their
product .K = 2 · 3 · 5 · · · pk . Let .P (k) denote the probability that two integers
a and b, selected randomly from the list .{1, 2, 3, . . . , K}, have no common
prime factor from⎛the list .{2, ⎞ 3, 5, . . . , pk }. Assuming the limits exist, show
∏ 1
. lim P (k) = 1 − 2 , where .P is the set of all primes.
k→∞ p
p∈P

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 15

P2007-5 Let A be an .n × n matrix such that .aij = 1 when .i /= j , and .aij = 0

⎡ ⎤
0 1 1 ··· 1
⎢1 0 1 · · · 1⎥
⎢ ⎥
⎢ ⎥
when .i = j . In other words, .A = ⎢1 1 0 · · · 1⎥. Find .A−1 . Hint: Using the matrix
⎢. . . . .⎥
⎣ .. .. .. . . .. ⎦
1 1 1 ··· 0
⎡ ⎤
111 ···
1
⎢1 1 1 1⎥
···
⎢ ⎥
⎢ 1⎥
···
.B = ⎢1 1 1 ⎥ may be helpful.
⎢. . . .. ⎥
..
⎣ .. .. .. .⎦.
1 1 1 ··· 1

P2007-6 Let g and h be non-commuting elements in a group of odd order.

Assuming g and h satisfy the relations .g 3 = e and .ghg −1 = h3 , determine the
order of h.
The 2007 Problem Solutions begin on page 95 in Chap. 30.
Chapter 8
2008 Problems

P2008-1 Let S be a set with a binary operation .∗ that is associative. Suppose that
for all x and y in S we have .x ∗ x ∗ x = x (i.e. .x 3 = x) and .x ∗ x ∗ y = y ∗ x ∗ x
(i.e. .x 2 y = yx 2 ). Show that for all x and y in S we have that .x ∗ y = y ∗ x.

P2008-2 Two friends agree to meet at the library, but each has forgotten the time
they were supposed to meet. Each remembers that they were supposed to meet
sometime between 1:00 pm and 5:00 pm. They each independently decide to go
to the library at a random time between 1:00 pm and 5:00 pm, wait for 30 min, and
leave if the other does not show up. What is the probability that they meet during
this 4-h period?

P2008-3 Suppose that two triangles have a common angle. Show that the sum of
the sines of the angles will be larger in that triangle where the difference of the
remaining two angles is smaller. Hint:
⎛ ⎞ ⎛ ⎞
θ +δ θ −δ
.2 sin cos = sin(θ ) + sin(δ).
2 2

P2008-4 The base of a solid object is the region bounded by the parabola .y = 12 x 2
and the line .y = 2; cross-sections of the object perpendicular to the y-axis are
semicircles. What is the volume of the object?

P2008-5 Define the sequence .{xn }∞

n=0 by .x0 = 0, .x1 = 1,

xn−1 + (n − 1)xn−2
xn =
. .
n
Determine . lim xn . Hint: Recalling that
n→∞

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 17

∞
⎲ (−1)k
. ln(1 + x) = x k+1 , x ∈ (−1, 1]
k+1
k=0

may be helpful.
⎛ ⎞
1 2n
P2008-6 Prove that . is an integer for all integers .n ≥ 1.
n+1 n
⎾ ⏋
1 −1
P2008-7 Find matrices B and C such that .B 3 + C 3 = .
0 5
The 2008 Problem Solutions begin on page 99 in Chap. 31.
Chapter 9
2009 Problems

P2009-1 Suppose that the points P and Q are randomly selected in the interval
[0, 2], so the segment from P to Q has length .|P Q| ≥ 0. What is the probability
.

Pr.[|P Q| ≤ 13 ]?

P2009-2 Show that the Maclaurin series of

x
f (x) =
.
1 − x − x2

⎲∞
is equal to . fn x n , where .f1 = 1, .f2 = 1, and .fn = fn−1 + fn−2 .
n=1

P2009-3 Let .T = {1, 2, 3, 4, 5, 6, 7, 8}. Let the set S be defined as follows.

S = {f |f : T → T is a bijection}
.

with binary operation function composition. Let .σ ∈ S. Suppose that .σ 3 is defined

as follows.

σ 3 (1) = 2
.

σ 3 (2) = 3
σ 3 (3) = 5
σ 3 (4) = 6
σ 3 (5) = 7
σ 3 (6) = 1

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 19

σ 3 (7) = 8
σ 3 (8) = 4

What is .σ ?
√
3 √ √
3 √
P2009-4 Show that . 2 + 5 + 2 − 5 = 1.

P2009-5 Suppose that n is a composite number, .n > 0, and .n /= 4. Show that

n|(n − 1)!.
.

P2009-6 Given .a, b ∈ R with .a < b, and a function .f : (a, b) → R, suppose that
f is increasing and satisfies the property that for all .λ ∈ (0, 1) and .x, y ∈ (a, b),

f (λx + (1 − λ)y) ≤ λf (x) + (1 − λ)f (y).

Prove that f is continuous on .(a, b).

P2009-7 Let .Z≥a be equal to the set .{x|x ∈ Z, x ≥ a}. It is known that there is a
one-to-one correspondence

F : Z≥0 × Z≥0 × Z≥0 → Z≥1 .

Write down a formula for F .

The 2009 Problem Solutions begin on page 103 in Chap. 32.
Chapter 10
2010 Problems

P2010-1 Suppose that you have a deck of n cards, divided into two stacks. One
stack contains p red suit cards and the other stack contains q black suit cards, with
.p + q = n. You completely shuffle the stacks together. How many red–black color

rearrangements of the cards are possible, assuming p and q are fixed?

P2010-2 Suppose that p is an odd prime and that .0 ≤ k ≤ p − 1. Prove that

(p − (k + 1))!k! ≡ (−1)k+1 (mod p).

P2010-3 Consider the following game. There are two players, player 1 and player
2. There is a (non-empty) pile of coins, each identical, on the gaming table. Player
1 acts first and must remove either 1, 2, or 3 coins. Player 2 acts next and must
remove either 1, 2, or 3 coins. The players continue taking turns in the manner
described until there are no coins left on the gaming table. The player who selects
the last coin is the loser. It is known that player 1 has a strategy that will guarantee a
win if the number of chips on the table is equivalent to 0, 2, or 3 modulo 4. Explain
what this strategy is, and prove that the strategy will guarantee the win for player 1.

P2010-4 Suppose that A and B are .n × n matrices with real entries with the
following two properties:

rank(A) + rank(B) = n
.

and

A + B = I,
.

where I is the .n × n identity matrix. Prove that .A2 = A, .B 2 = B, and .AB = BA.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 21

P2010-5 Evaluate the following limit.

sin(arctan(x)) − tan(arcsin(x))
. lim .
x→0 arcsin(tan(x)) − arctan(sin(x))

P2010-6 Consider the lattice .Z × Z. Imagine that there is a point at the origin.
The point will move in one of four directions in the lattice: up (the y coordinate is
increased by one unit), right (the x coordinate is increased by one unit), left (the
x coordinate is reduced by one unit), or down (the y coordinate is reduced by one
unit). The probability that the point will move up is .1/4, the probability that it will
move down is .1/8, the probability that it will move right is .1/16, and the probability
that the point will move left is .9/16. Suppose that we keep track of the movement
of the point for 5 rounds. What is the probability that it winds up in the open sector
(not including the points along the line) in the first quadrant defined by the lines
.y = 3x and .y = x?
1
2

P2010-7 A rational number can be written . pq where p and q are integers, .q > 0,
and p and q have no common factors. Let the function .f (x) be defined as follows:
If x is irrational, then .f (x) = 0. If x is rational, then .f (x) = q1 . Prove that this
function is continuous at every irrational.
The 2010 Problem Solutions begin on page 109 in Chap. 33.
Chapter 11
2011 Problems

P2011-1 Suppose that children are playing a game together, where an even
number, .n ≥ 6, of them are seated in a circle, facing the center. One of these
children is designated as child number 1. The child to the right of child number
1 is designated as child number 2, and so on around the circle in a counterclockwise
fashion. There is another child in the center of the circle, who is able to choose a
positive integer i (which can be larger than n), known as the elimination parameter.
Starting with child number 1, the child in the center of the circle counts off the
first i children in a counterclockwise fashion (moving around the circle more than
once if necessary). The child following the ith child leaves the circle and the game.
The pattern continues, counting another i children starting with the next child in the
circle after the child who has just left, until only one child remains in the circle,
and that child is declared the winner. One can make a list of the children as they
leave the game, establishing an order of elimination. Is there some n so that for any
ordered list of .n − 1 children from the circle, the child in the center can choose an
elimination parameter i that realizes this list as an elimination order?

P2011-2 Suppose that we have a sheet of cardboard in the shape of an equilateral

triangle of side length 1. You cut the corners off of the triangle by making a cut of
length x units perpendicular to each side. At that point, the sides are then folded up
and the box is formed. Show that the volume of the box is maximized when the area
of the base of the box is positive and equal to the total area of the sides of the box.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 23

x
1

P2011-3 Let n be a positive integer. Consider, for every integer i with .1 ≤ i ≤ n,

the set .Ci of real degree i polynomials with leading coefficient ./= 0 and constant
term .= 0, defined as follows:

Ci = {g(x) = a0 x i + a1 x i−1 + · · · + ai−1 x | a0 /= 0, a1 , . . . , ai−1 ∈ R}.

Let .C n = C1 ∪ C2 ∪ . . . ∪ Cn . Note that in particular for .i = 1, .C 1 = C1 is this set

of linear polynomials, .{g(x) = a0 x|a0 /= 0}.
Consider the following claim, depending on n: If .f : R2 → R is a function such
that for every .g ∈ C n

. lim f (x, g(x)) = 0,

x→0

then

. lim f (x, y) = 0.
(x,y)→(0,0)

Find some positive integer n for which the claim is true, and prove the claim, or show
that the claim is false for all positive integers n by stating a specific counterexample
.f (x, y) for each n.

P2011-4 Let b be a positive integer greater than 1. If x is a number, then there is an

expansion of x with base b:
⎲
x=
. ai bi 0 ≤ ai < b.
i∈Z

Let n be a positive integer such that .(n, b) = 1. Show that the period of the
expansion of .1/n is the smallest positive integer r such that .br ≡ 1 mod n.
11 2011 Problems 25

P2011-5 Suppose that one has three pegs and two different stacks of n disks. Each
stack of disks sits on a different peg (so there is one peg with no disks on it). The
disks in one stack are painted red, and the disks in the other stack are painted blue.
Each disk in the red stack has a different radius. The disk at the top of the red stack
has the smallest radius, the second topmost disk has the second smallest radius, and
so on all the way down to the bottom of the stack (which will be the disk with
the largest radius). For each disk in the red stack, there is a disk in the blue stack
identical to it except for color. The disks in the blue stack are arranged in exactly the
same way as the disks in the red stack. The goal is to exchange the stacks of disks.
There are some rules which limit possible moves. One can only move one disk at a
time, and one cannot stack a larger disk upon a smaller disk. One is allowed to stack
two disks which have the same size, but are different colors. Devise an algorithm
which will exchange the position of the two stacks, and calculate the number of
moves that your algorithm requires to perform the exchange.

P2011-6 Suppose that .f : R → R is a differentiable function and that .|f ' (x)| ≤
1/2 for every x. A number x is a fixed point for .f (x) means that .f (x) = x. Let
.x1 be an arbitrary real number. Suppose that we define a sequence as follows. Let

.x2 = f (x1 ); let .x3 = f (x2 ), and in general, let .xn = f (xn−1 ). Show that the

following limit exists:

. lim xn .
n→∞

Also, show that the limit is a fixed point for the function .f (x).

P2011-7 Let A and B be two sets. A function

f :A→B
.

is a one-to-one correspondence means that f is both one-to-one and onto. Show that
there is a one-to-one correspondence between .R and .R \ {0}.

P2011-8 For an .n × n matrix A with complex entries, a vector .v ∈ Cn is an

eigenvector for A means that .v /= 0→ and .Av = λv for some .λ ∈ C. For .n ≥ 2,
characterize the eigenvectors of the following .n × n matrix:
⎡ ⎤
1 23 ··· n
⎢ n 12 ··· n − 1⎥
⎢ ⎥
⎢ ··· n − 2⎥
.A = ⎢ n − 1 n1 ⎥.
⎢ . .. .. .. .. ⎥
⎣ .. . . . . ⎦
2 3 4 ··· 1

The 2011 Problem Solutions begin on page 113 in Chap. 34.

Chapter 12
2012 Problems

P2012-1 Show that .n2 divides .(n + 1)n − 1 for any positive integer n.

P2012-2 How many zeros are at the end of .213 !?

P2012-3 Let .p(x) = an x n + · · · + a2 x 2 + a1 x + a0 be a polynomial of degree

.n ≥ 2, with integer coefficients:

(a) Show that if .a0 , a1 , an , and .a2 + · · · + an are odd integers, then .p(x) has no
rational root.
(b) Give examples to show that the conclusion from part (a) may not be true if any
of .a0 , .a1 , .an , or .a2 + · · · + an is even.

P2012-4 Let A be an .n × n matrix whose diagonal entries are all equal to the
same real number .α ∈ R and all other entries are equal to .β ∈ R. Show that A is
diagonalizable, and compute the determinant of A.

P2012-5 Let G be a group of order 26. Show that if G has a normal subgroup of
order 2, then G is a cyclic group.

P2012-6 Show that for any positive integer k, the following is an irrational number:
∞
⎲ 1
. .
(n !)k
n=0

P2012-7 Let .f : [0, 1] → R be a continuous function. Show that

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 27

⎰ 1 1
. lim nx(1 − x 2 )n f (x) dx = f (0).
n→∞ 0 2

P2012-8 Recall that a function .f (x, y) is harmonic in an open subset .O of

the plane means that it is twice continuously differentiable in .O and .fxx (x, y) +
fyy (x, y) = 0 for all .(x, y) in .O. Let .R be the region in the plane given by

R = {(x, y) ∈ R2 : x 2 + (y + 1)2 ≤ 9 and x 2 + (y − 1)2 ≥ 1}.

Show that if f is harmonic in an open disk .O containing .R, then

⎰⎰
. f (x, y) dxdy = 9πf (0, −1) − πf (0, 1).
R

The 2012 Problem Solutions begin on page 121 in Chap. 35.

Chapter 13
2013 Problems

P2013-1 Define
√ a sequence .(sn ) recursively as follows: Let .s1 = 1, and for .n ≥ 1,
let .sn+1 = 1 + sn . Prove that .(sn ) converges, and then find the limit.

P2013-2 Let .C be a non-empty collection (possibly infinite) of compact subsets of

R:
.

⋂
(a) Prove that .K = C is a compact set.
C∈C
(b) Give an example that illustrates that the union of a family of compact sets need
not be compact.

P2013-3 Assume A and B are two sets with m and n elements, respectively:

(a) How many one-to-one functions are there from A to B?

(b) How many one-to-one and onto functions are there from A to B?

P2013-4 Let p and q be distinct prime numbers. Find the number of generators of
the group .Zpq .

P2013-5 Let G be a group and H a subgroup of G with index .(G : H ) = 2. Prove

that H is a normal subgroup of G.

P2013-6 The Fibonacci numbers are defined as

f1 = f2 = 1
.

and

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 29

. fn+1 = fn + fn−1

for .n ≥ 2:

(a) List .f1 , f2 , . . . , f7 .

(b) Illustrate, using the list from (a), that .f2n+1 = fn+1
2 + f 2 for .n = 1, 2, 3.
n
(c) Prove that .f2n+1 = fn+1 + fn for all .n ∈ N.
2 2

P2013-7 Let .a, b, m, M be real numbers with .0 < m ≤ a ≤ b ≤ M. Prove that

√ √
2 mM 2 ab
. ≤ .
m+M a+b

P2013-8 A soccer ball is stitched together using white hexagons and black
pentagons. Each pentagon borders five hexagons. Each hexagon borders three other
hexagons and three pentagons. Each vertex is of valence 3 (meaning that at each
corner of a hexagon or pentagon, exactly three hexagons or pentagons meet). How
many hexagons and how many pentagons are needed to make a soccer ball? Hint:
Euler’s formula for convex polyhedra states that .V − E + F = 2, where V is the
number of vertices, E is the number of edges (i.e., the line adjoining two vertices),
and F is the number of faces (hexagons or pentagons).
The 2013 Problem Solutions begin on page 131 in Chap. 36.
Chapter 14
2014 Problems

P2014-1 Let .a > 0, and define the following function:

√ √
3
a3x − a a2x
.f (x) = √ . (14.1)
4
a − ax 3

(a) Calculate these limits.

. lim f (x) =
x→0+
lim f (x) =
x→a
lim f (x) =
x→+∞
.

(b) Find the maximum value of .f (x) on its domain.

P2014-2 Let f be a function with domain .(0, ∞) satisfying:

• .f (x) = f (x 2 ) for all .x > 0.

• . lim f (x) = lim f (x) = f (1).
x→0+ x→+∞

Show that .f (x) is a constant function on .(0, ∞).

P2014-3 Let V be a corner of a right-angled box and let x, y, z be the angles formed
by the long diagonal and the face diagonals starting at V . For

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 31

⎡ ⎤
sin x sin y sin z
.A = ⎣ sin z sin x sin y ⎦

sin y sin z sin x

show that .|det(A)| ≤ 1.

P2014-4 Let .f (t) be a real-valued integrable function on .[0, 1], so that both sides
of the following equation are continuous functions of x:
⎰ x
2x − 1 =
. f (t)dt.
0

Prove that if .f (t) ≤ 1 for .0 ≤ t ≤ 1, then there exists a unique solution .x ∈ [0, 1]
of the equation.

P2014-5 Let ABCD be a rectangle. The bisector of the angle ACB intersects AB
at point M and divides the rectangle ABCD into two regions: the triangle MBC
with area s and the convex quadrilateral MADC with area t:

(a) Determine the dimensions of the rectangle ABCD in terms of s and t.

(b) In the case .t = 4s, what is the ratio .AB/BC?

P2014-6 In a badly overcrowded preschool, every child is either left-handed or

right-handed, either blue-eyed or brown-eyed, and either a boy or a girl. Exactly half
of the children are girls, exactly half of the children are left-handed, and exactly one
fourth of the children are both. There are twenty-six children who are brown-eyed.
Nine of those twenty-six are right-handed boys. Two children are right-handed boys
with blue eyes. Thirteen children are both left-handed and brown-eyed. Five of these
thirteen are girls:

(a) How many students does the preschool have?

(b) How many girls are right-handed and blue-eyed?

P2014-7 Let .n > 1 be an integer. Let .(G, ·) be a group, with an identity element e
and an element .a ∈ G with .a /= e and .a n = e. Let .(H, ∗) be a group, let .f : G → H
be an arbitrary function, and then define .F : G → H by

F (x) = f (x) ∗ f (a · x) ∗ f (a 2 · x) ∗ . . . ∗ f (a n−1 · x).

(a) Show that if .f (G) is a subset of some Abelian subgroup of H , then F is not a
one-to-one function.
14 2014 Problems 33

(b) Let .(H, ∗) be the symmetric group .(S3 , ◦) (the six-element group of permu-
tations of three objects). Give an example of .(G, ·), n, and a as above, and a
function .f : G → H , so that the expression F is a one-to-one function.

P2014-8 Determine whether the following sum of real cube roots is rational or
irrational.
/ / / /
3 847 3 847
. 6+ + 6− .
27 27

The 2014 Problem Solutions begin on page 137 in Chap. 37.

Chapter 15
2015 Problems

P2015-1 An integer n has a super-3 representation means that there is a positive

integer m, a sequence of distinct nonnegative integers .p1 , . . . , pm , and a sequence
.a1 , . . . , am where each .ak is .±1, so that

⎲
m
n=
. ak · 3pk = a1 · 3p1 + · · · + am · 3pm .
k=1

For instance, the integer 8 has the super-3 representation .8 = 32 −30 , and the integer
.−11 has the super-3 representation .−11 = −3 − 3 + 3 . The number 0 has the
2 1 0

empty super-3 representation, i.e., where .m = 0 and the sum has no terms:

(a) Give a super-3 representation of 2015.

(b) Prove that every integer n has a super-3 representation.

P2015-2 Prove that, for every positive integer n,

⎲ n−k ⎛ ⎞⎛
n ⎲ ⎞
n n − k k+i
. 2 = 5n .
k i
k=0 i=0

P2015-3 Three pairwise perpendicular line segments .AB, .CD, and .EF have
endpoints all on a sphere of unknown radius, and intersect inside the sphere at a
point X. Given the lengths .AX = 1, .CX = 2, .EX = 3, and .BX = 4, determine,
with proof, the volume of the octahedron with vertices A, B, C, D, E, and F .

P2015-4 You are standing in a room, which we will call .Σ, which contains eight
light switches numbered 1 through 8, all in the off position. On the other side of the

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 35

door is another room, .Λ, which contains eight lights, labeled A through H . Each
switch controls exactly one light. Your goal is to determine which switches control
which lights. You do this by making a number of trials, which consist of putting
some set of switches in room .Σ in the on position, and then entering room .Λ to
discover which lights are on. For instance, one trial might be to turn switches 1 and
3 to the on position (and all others off) and observe which lights in room .Λ are on
(perhaps lights D and H , though of course you cannot know this ahead of time):

(a) Give a strategy for finding out which switches control which lights using the
smallest possible number of trials. (For this part, you do not need to prove that
the number of trials you use is minimal.)
(b) Prove that your strategy uses the minimal number of trials; that is, prove that
there is no strategy that determines which switches control which lights using
fewer trials.

P2015-5 A function .f : R → R is smooth means that all of its derivatives (first,

second, third, etc.) exist everywhere:

(a) Suppose .f : R → R is a smooth function with infinitely many zeros in the

interval .[0, 1]. Show that there is some .x ∈ [0, 1] such that .f (n) (x) = 0 for
every integer .n ≥ 0 (that is, such that f and all its derivatives vanish at x).
(b) Give an example of such a function f which is nonconstant on every interval.
(You need not prove that your function works.)

P2015-6 Recall that a graph is planar means that it can be drawn on the plane in
such a way that no two of its edges cross. Further recall that a graph is c-colorable
(for a positive integer c) means that its vertices can be colored with c colors in such
a way that no two adjacent vertices have the same color. The famous Four Color
Theorem says that every planar graph can be 4-colored.
A graph is k-color-planar means that it can be drawn on the plane, and its edges
colored with k colors, in such a way that no two edges with the same color cross.
Thus, a 1-color-planar graph is just a planar graph.
Shown below are two graphs illustrating these definitions. Graph G is planar and its
vertices can be 3-colored as shown, with the three colors represented by .•, .◦, and
.⊗. The graph H is 2-color-planar (but not planar), with the two colors represented

by thin and thick edges.

15 2015 Problems 37

H
G
⊗

Prove that, for every positive integer k, there is a positive integer .ck such that every
k-color-planar graph is .ck -colorable. Hint: You may use the Four Color Theorem in
your proof if you wish.

P2015-7 Let S be a non-empty, finite set and .∗ : S → S be a binary operation on

S. Suppose that .∗ satisfies the following two conditions:

• .∗ is associative; that is, for any .a, b, c ∈ S, .(a ∗ b) ∗ c = a ∗ (b ∗ c).

• For any .a, b ∈ S, .a ∗ (a ∗ (b ∗ a)) = b:

(a) Prove that .∗ is commutative; that is, that for any .a, b ∈ S, .a ∗ b = b ∗ a.
(b) Prove that .|S|, the size of S, is a power of 3.

P2015-8 Let .f : R → R be twice differentiable on .[0, 1], and suppose that

.f (0) = f ' (0) = 0 and .f (1) = 1. Prove that there is some .a ∈ (0, 1) such that
' '' 9
.f (a)f (a) = .
8
The 2015 Problem Solutions begin on page 145 in Chap. 38.
Chapter 16
2016 Problems

P2016-1 Assume we are looking at numbers in normal decimal representations,

and consider the set

A = {1, 11, 111, 1111, . . .} = {x ∈ Z | x consists entirely of 1s}.

For all .a ∈ A, define .n(a) to be the number of 1s in a’s representation. (For example,
.n(1) = 1 and .n(111) = 3.) Prove or disprove the following statement: The set
.{a ∈ A | n(a) divides a} is infinite.

P2016-2 Let .f : Z → Z be a function with all of the following properties:

(i) .f (2) = 2.
(ii) .f (mn) = f (m)f (n) for all .m, n ∈ Z.
(iii) If .m > n, then .f (m) > f (n).

Conjecture what all the possible candidates for f are, and then prove your
conjecture.

P2016-3 Arrange eight points in 3-space so that each of the 56 possible triplets
of points determined forms an isosceles triangle. Prove your arrangement works.
(Note: Some triangles may be degenerate!)

P2016-4 Let A be a subset of .R, and let .f, g : A → A be two continuous functions.
The function f is homotopic to g means that there is a continuous function .h :
A × [0, 1] → A such that .h(a, 0) = f (a) and .h(a, 1) = g(a) for all a in A:
(This function is sometimes called a deformation function. It may be helpful to
think of .[0, 1] as being a “slider” that continuously morphs the function f into the
function g.)

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 39

(a) Let .A = [0, 1] ⊂ R, and define .f, g : A → A by .f (a) = a 2 and .g(a) = a 3 .

Find a function h demonstrating that f is homotopic to g.
(b) Recall that an equivalence relation .∼ on a set X requires three properties:
• Reflexivity: .∀ x ∈ X, .x ∼ x.
• Symmetry: .∀ x, y ∈ X, if .x ∼ y, then .y ∼ x.
• Transitivity: .∀ x, y, z ∈ X, if .x ∼ y and .y ∼ z, then .x ∼ z.
Prove that for any subset A of .R, being homotopic forms an equivalence relation
on the set .C(A) of all continuous functions .f : A → A.

P2016-5 The Fibonacci numbers .fn (.n = 0, 1, 2, . . . ) are defined recursively by

f0 = 0, .f1 = 1, and .fn = fn−2 + fn−1 for .n ≥ 2:
.

(a) Find a linear transformation .T : R2 → R2 such that .T (fn , fn+1 ) =

(fn+1 , fn+2 ) for all .n ≥ 0, and then state and prove a conjecture for what
.T (0, 1) equals for all .n ≥ 1.
n

(b) Find the matrix A of T with respect to the standard basis for .R2 , and then find
the eigenvalues of A.
(c) Find a non-recursive expression for .fn for all .n ≥ 1.

P2016-6 Let A be an open subset of the real numbers and .f : A → R be a

differentiable function such that .f ' (a) = 0 for all .a ∈ A:

(a) Prove that if .A = R, then f must be a constant function. (Note: Do not use
antiderivatives in your proof, as the proof that antiderivatives are unique up to a
constant depends on this statement.)
(b) Prove or disprove: If .A /= R, then f must be a constant function.

P2016-7 Let G be a set and .∗ : G × G → G be a binary operation. G, together

with the operation .∗, is a group means that the following conditions are satisfied:

• Associativity: .∀ a, b, c ∈ G, .(a ∗ b) ∗ c = a ∗ (b ∗ c).

• Identity: .∃e ∈ G such that .∀ g ∈ G, .e ∗ g = g ∗ e = g.
• Inverse: .∀ a ∈ G, .∃ a −1 ∈ G such that .a ∗ a −1 = a −1 ∗ a = e.

For purposes of shorthand, the symbol .∗ is often omitted (i.e., we write .a ∗ b as

simply ab and .a ∗ a ∗ a as simply .a 3 ).
Assume that G is a group which has the following additional properties:

(i) For all .g, h ∈ G, (gh)3 = g 3 h3 .

(ii) There are no elements in G of order 3 (i.e., .∄ g ∈ G such that .g 3 = e).
16 2016 Problems 41

For any group G with properties (i) and (ii) listed above:

(a) Prove that for all .a, b ∈ G, if .a /= b, then .a 3 /= b3 .

(b) Prove that the map .φ : G → G defined by .φ(g) = g 3 is a bijection if G is
finite.

P2016-8 Remark from the problem authors : The length of this puzzle problem is
for clarity, not necessarily its difficulty.
Assume elections are conducted by voters who place all the candidates in rank order.
Under these conditions there are several possible voting methods available:

• Plurality—the candidate with the most first place votes wins

• Plurality with elimination—the following process is repeated until there is a
winner:
– If there is a candidate that has more than half the first place votes, he/she wins.
– Otherwise, eliminate the candidate(s) with the fewest first place votes. Reorder
the voters’ preferences by moving other candidates up in the ranks to fill the
vacancies.
• Borda count—Candidates receive points based upon each voter’s ranking. A
last-place vote earns a candidate one point, a second-to-last-place vote earns a
candidate two points, a third-to-last-place vote earns a candidate three points,
and so forth, with a first place vote earning a candidate n points, where n is the
total number of candidates. The candidate with the most total points wins.
• Pairwise Comparison—Candidates receive points based upon their performance
in a round-robin style analysis. Each candidate is paired with each other
candidate. (For example, if there were four candidates, there would be six
possible matchups.) For each matchup, count the number of voters who prefer
one candidate over the other. The winner of the matchup (more voter preferences)
receives a point; if there is a tie, both candidates receive half a point. The
candidate with the most total points wins.
• Survivor—The following process is repeated until there is a winner:
– If there is only one candidate remaining, he/she wins.
– Otherwise, eliminate the candidate(s) with the most last-place votes. Reorder
the voters’ preferences by moving other candidates up in the ranks to fill the
vacancies.

Example Consider the following sample preference table for nine voters in an
election with three candidates: X, Y , and Z:
42 16 2016 Problems

Rank/Number of Votes 4 3 2
First X Y Z
. .
Second Y Z Y
Third Z X X

That is, four voters rank X as their first choice, Y as their second choice, and Z as
their third choice; three voters would pick Y first, Z second and X third; and two
voters would choose Z first, Y second, and X third. Using this table of preferences,
the winners for the various voting methods would be:

• Plurality—X wins with the most first place votes (4 of them).

• Plurality with elimination—No candidate has more than half the first place votes,
so we eliminate the candidate(s) with the fewest. Candidate Z is eliminated. The
new preference table is

Rank/Number of Votes 4 3 2
. First X Y Y .
Second Y X X

Now Y wins with more than half the first place votes.
• Borda count—The point totals are as follows:
– X: 4(3)+3(1)+2(1)=17
– Y : 4(2)+3(3)+2(2)=21
– Z: 4(1)+3(2)+2(3)=16
Therefore, Y wins.
• Pairwise comparison—The matchups are as follows:
– X vs. Y : Y wins 5 to 4 because 5 voters like Y better than X; Y gets a point.
– X vs. Z: Z wins 5 to 4 because 5 voters like Z better than X; Z gets a point.
– Y vs. Z: Y wins 7 to 2 because 7 voters like Y better than Z; Y gets a point.
Therefore, Y wins.
• Survivor—We eliminate the candidate(s) with the most last-place votes. Candi-
date X is eliminated. The new preference table is

Rank/Number of Votes 4 3 2
. First Y Z Y .
Second Z Y Z

Now Z is eliminated, leaving only Y remaining; therefore, Y wins.

Challenge: Construct a preference table for an election with five candidates—A, B,

C, D, and E—such that A wins via plurality, B wins via plurality with elimination,
16 2016 Problems 43

C wins via Borda count, D wins via pairwise comparison, and E wins via survivor.
You may use any number of voters you would like.
The 2016 Problem Solutions begin on page 151 in Chap. 39.
Chapter 17
2017 Problems

P2017-1 In constructible geometry, one constructs points, lines, and circles from
given points, lines, and circles, using an unmarked straightedge and compass:

• The straightedge draws a line between points already given, which includes the
line segment connecting them; the line may extend as far beyond either point as
desired. New points are created where the line intersects already existing lines or
circles.
• The compass draws a circle (or a circular arc) centered on a given point with a
radius extending to another point from the center. Again, new points are created
where the circle intersects other circles or lines:

(a) Use a straightedge and compass to construct the midpoint M of the line segment
.AB given below. (Show your construction and label M. Do not erase any

intermediate steps.)

(b) Use a straightedge and compass to construct a square ABCD with the line
segment .AB given below as one of its sides. (Show your construction and labels
C and D. Do not erase any intermediate steps.)

(c) Given an isosceles right triangle ABC, one can use a compass to construct
lunes as follows. First, one semicircle is formed with the line segment .AC as its
diameter. Two other semicircles are formed with .AB and .BC as their diameters.
The lunes are the shaded shapes in the ﬁgure below.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 45

Use a straightedge and compass to construct a square whose area is equal to the
area of one of the lunes on the diagram given. Justify how you know the square
you construct has the appropriate area.

P2017-2 Consider the following series

∞
⎲
. (an )n ,
n=1

where
⌠
1
+ 312 + · · · + 1
if n is odd,
an =
. 31 3n
| sin(n) cos(n)| if n is even.

Make a conjecture as to whether the series converges or not, and then prove your
conjecture.

P2017-3 Two students, Joe and Frank, are each asked to independently select a
number at random from the interval [0,1] using a uniform probability distribution
(where the probability of selecting a number from a subinterval .[a, b] is equal to
the length .b − a). If a denotes the number chosen by Joe and b denotes the number
chosen by Frank, what is the probability that the quadratic equation .x 2 + ax + b = 0
has at least one real root?

P2017-4 A multiplicative magic square is an .n × n square array of numbers

consisting of .n2 distinct positive integers (not necessarily consecutive) arranged
such that the product of the n numbers in any of the n rows, n columns, or two
main diagonal lines is always the same number. Call this common product the magic
product:
17 2017 Problems 47

(a) Show that the magic product of a .3 × 3 multiplicative magic square must be a
perfect cube.
(b) Find an example of a .3 × 3 multiplicative magic square whose magic product
is minimal. Explain how you know this magic product is minimal.

P2017-5 For any positive integer n, let .s(n) be the sum of the ﬁrst n terms of the
sequence

0, 1, 1, 2, 2, 3, 3, 4, 4, . . . , k, k, k + 1, k + 1, . . .
.

(a) Find a formula for .s(n). (Note: Your ﬁnal formula should not have “.. . . ” in it.)
(b) Suppose that m and n are any two positive integers with .m > n. Prove that
.s(m + n) − s(m − n) = mn.

P2017-6 Suppose that A is an .n × n matrix such that every entry of A is .±1. Show
that the determinant of A is divisible by .2n−1 .

P2017-7 A robot is programmed to shufﬂe cards in such a way so that it always

rearranges cards in the same way relative to the order in which the cards are given
to it. The thirteen hearts arranged in the order

A, 2, 3, 4, 5, 6, 7, 8, 9, 10, J, Q, K
.

are given to the robot, shuffled, and then the shuffled cards are given back to the
robot and shuffled again. This process is repeated until the cards have been shuffled
a total of 7 times. At this point the order of the cards is

2, 4, 6, 8, 10, Q, A, K, J, 9, 7, 5, 3.
.

What was the order of the cards after the ﬁrst shufﬂe?

P2017-8 Recall that a subset A of .R is closed in .R means that the complement

R \ A is equal to a union of any number of open intervals (possibly inﬁnitely many).
.

Suppose A is a non-empty, closed subset of .R such that for each .a ∈ A, every open
interval that contains a also contains another element of A. Show that A must be
uncountable.
The 2017 Problem Solutions begin on page 159 in Chap. 40.
Chapter 18
2018 Problems

P2018-1 Show that

. sin(x) sin(2x) · · · sin(nx) /= 1

for every real number x and any positive integer .n ≥ 2.

P2018-2 Determine the smallest natural number n such that

1 1 1
. √ +√ √ + ··· + √ √ ≥ 100.
1+ 2 2+ 3 n+ n+1

P2018-3 Suppose .a1 , .a2 , . . . , .an are strictly positive real numbers and

a1x + a2x + · · · + anx ≥ n,

for every real number x. Prove that .a1 a2 · · · an = 1.

P2018-4 Show that if .x + y + z > 0, then

⎡ ⎤
x zy
. det ⎣ y x z ⎦ ≥ 0.

zyx

P2018-5 Consider the following sequence defined recursively:

1
x1 =
. , xk+1 = xk2 + xk , k ≥ 1.
2

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 49

Find the integer part of .S100 , where:

1 1 1
S100 =
. + + ··· + .
x1 + 1 x2 + 1 x100 + 1

P2018-6 Consider a semicircle and AB its diameter. Pick two arbitrary points D
and E on the semicircle such that the segments AD and BE intersect at M in the
interior of the semicircle. Prove that:

|AM| · |AD| + |BM| · |BE| = |AB|2 .

P2018-7 Suppose a fair six-sided die is rolled, and let X denote the outcome of
the die roll. Suppose a fair coin is flipped until X heads are observed. Compute the
expected value of the number of flips.

P2018-8 A town has n inhabitants who like to form clubs. They want to form clubs
so that every pair of clubs should share a member, but no three clubs should share
a member. What is the maximum number of clubs they can form? Illustrate with an
example.
The 2018 Problem Solutions begin on page 167 in Chap. 41.
Chapter 19
2019 Problems

P2019-1 Provide counterexamples to each of the following statements:

(a) If both .f (x) and .g(x) are continuous and monotone on .R, then .f (x) + g(x) is
continuous and monotone on .R.
(b) If a function .f (x) is not bounded in any neighborhood of a point a, then either
. lim |f (x)| = ∞ or . lim |f (x)| = ∞.
x→a + x→a −
(c) A function cannot be continuous at only one point in its domain and discontin-
uous everywhere else.
(d) If a function is differentiable and increasing on an interval .(a, b), then its
derivative is positive on the interval .(a, b).
(e) If .f (x) is a function with an antiderivative .F (x) that is defined at both a and b,
⎰ b
then . f (x)dx = F (b) − F (a).
a

P2019-2 Let .f (x) = x 2 /4, and consider the set of all right triangles in the
plane whose right angle vertex lies at the origin and whose other two vertices lie
somewhere else on the graph of .y = f (x). Conjecture on the existence and location
of a point other than the origin through which all such triangles must pass, and then
prove your conjecture.

P2019-3 Two-Face and his henchmen are once again on a crime spree terrorizing
the good people of Gotham City. However, rather than flipping a coin to decide his
behavior, Two-Face has decided to adopt a more deterministic approach. When Two-
Face’s gang crashes a party with 60 guests in attendance, they first line everyone up
against the wall. Then, the first henchman walks down the line and takes .$10 from
every guest. Next, a second henchman walks down the line giving .$10 to the second,
fourth, sixth, etc., party guests. This process is repeated with a third henchman
taking .$10 from the third guest and every third guest thereafter, then with a fourth

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 51

henchman giving .$10 to the fourth guest and every fourth guest thereafter, and so
on, until finally the sixtieth henchman gives .$10 to only the sixtieth (last) person in
line:

(a) How many party guests made money? How many lost money? How many broke
even?
(b) Which party guest(s) made the most money?
(c) Which party guest(s) lost the most money?
(d) How much money did Two-Face and his henchmen make?

P2019-4 A fair coin is a coin that will produce a result of either heads (H) or tails
(T) when flipped with equal probability. Suppose that you start flipping a fair coin
and record the sequence of H/T results:

(a) What is the expected number of flips needed to get your first tail (T)?
(b) What is the expected number of flips needed to achieve your first string of heads
followed by tails (HT)?
(c) What is the expected number of flips needed to achieve your first string of heads
followed by heads (HH)?

⎾ ⏋
1 2019
P2019-5 Let .A = . State a conjecture on the four entries of the matrix
0 1
.A for any .n ∈ Z, and then prove your conjecture.
n

P2019-6 Let .(sn )∞ ∞ ∞

n=1 , .(tn )n=1 , and .(un )n=1 be sequences with the following
properties:

• .(sn ) is monotone decreasing.

• .(tn ) is monotone increasing.
• .sn ≥ un ≥ tn for every .n ∈ N.

For each of the three sequences .(sn )∞ ∞ ∞

n=1 , .(tn )n=1 , and .(un )n=1 , conjecture that the
sequence must converge must not converge, or not enough information is given to
determine convergence. Then prove your conjectures.

P2019-7 Suppose that G is a group. A subset S of G is a set of generators for G

means that every element of G can be written as a finite product of elements in S
and/or their inverses. Suppose that G is a group with identity e which has .{x, y} as
a set of generators, where the generators x and y satisfy the following relations:

• .x 2 = e.
• .y 4 = e.
• .xyxyxy = e.
19 2019 Problems 53

Determine the maximum number of elements that G can contain. Justify your
answer.

P2019-8 Suppose that T is the triangular pyramid with vertices at .(0, 0, 0),
(12, 0, 0), .(0, 8, 0), and .(0, 0, 24). What is the maximum volume that a rectangular
.

prism R which has one vertex at .(0, 0, 0) and which is inscribed in T can have?
The 2019 Problem Solutions begin on page 173 in Chap. 42.
Chapter 20
2020 Problems

The year 2020 had enough problems.

In the spring of 2020, the Indiana Section of the MAA was scheduled to hold its
annual spring meeting in conjunction with an interdisciplinary, multi-organization
conference hosted by IUPUI to celebrate the 150th anniversary of the founding
of Indiana University in Bloomington. Unfortunately, the COVID-19 pandemic
erupted just a few months before the conference was scheduled, and the State of
Indiana declared a public health emergency and a stay-at-home order. The larger
conference was canceled, so the Section meeting was canceled, and consequently,
the 2020 edition of the Indiana College Mathematics Competition was canceled.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 55

⎾ ⏋
1 2021
P2021-1 Let .A = .
0 1

(a) Find .A2021 .

(b) Find a .2 × 2 matrix B so that .B 2021 = A.

P2021-2 The polynomial .p(x) = 16x 4 − 32x 3 − 104x 2 + 122x + 232 has an
interesting property. There is a line .y = mx + b that is tangent to the graph of p in
two places. Find the line. Hint: Consider the polynomial .q(x) = p(x) − (mx + b)
and how it might factor.

P2021-3 A three-digit positive integer n is exactly five times the product of its
digits.

(a) Show that the digits of n must all be odd.

(b) Find n.

P2021-4 Consider the points .A(3, 4, 1), .B(5, 2, 9), and .C(1, 6, 5) in .R3 . Show that
these points are the vertices of a cube.

P2021-5 Let .S = {1, 2, 3, 4, 5, 6, 7, 8}. A partition P of S into two-element subsets

{{x1 , y1 }, {x2 , y2 }, {x3 , y3 }, {x4 , y4 }} has each .xi and each .yi a value from S, with
.

all values used precisely once.

(a) How many partitions of S into four two-element subsets are there?

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 57

⎲
4
(b) For a given partition, define its value by .V (P ) = xi yi . An integer n is
i=1
achievable means that there is a partition P whose value .V (P ) is n. Find the
minimum and maximum achievable values.

P2021-6 Let .xyz be a three-digit number, made with digits x, y, and z. That is,
xyz = 100x + 10y + z.
.

(a) Find digits a, b, and c, not necessarily distinct, for which .abc + cab − bca =
608.
(b) Show that there are no values of a, b, and c that satisfy .abc + cab − bca = 707.

P2021-7 Polynomial p has nonnegative integer coefficients and satisfies .p(1) = 21

and .p(11) = 2021. What is .p(10)?

P2021-8 In a “KenKen” puzzle, the numbers in each heavily outlined set of

squares, called cages, must combine (in any order) to produce the target number
in the top corner of the cage using the mathematical operation indicated. A number
can be repeated within a cage as long as it is not in the same row or column. In the
.5 × 5 puzzle, each of the digits 1 through 5 must appear precisely once in each row

and column. Solve the .5 × 5 KenKen below.

55, + 288, ×

The 2021 Problem Solutions begin on page 181 in Chap. 43.

Chapter 22
2022 Problems

P2022-1 Consider the four lines .l1 : x = −3, .l2 : x = 1, .l3 : y = 2, and
l4 : y = −4. If A is some point in the plane, suppose each of the segments from A
.

to the lines meets perpendicularly at B, C, D, and E, respectively (i.e., B is on .l1 ,

C on .l2 , etc.). Consider the locus of all points A where

|AB||AC| = |AD||AE|.
.

Find an equation describing this locus, and specify as much as possible the type of
plane curve it is.

P2022-2 Find two nonzero functions .f (x) and .g(x) so that .f ' (x) /= 0, .g ' (x) /= 0,
and
d
. (f (x)g(x)) = f ' (x)g ' (x).
dx

1
P2022-3 Consider the function .f (x) = .
1 − e−1/x

(a) Assuming .x > 0, find .f ' (x).

⎰ 1
e−1/x
(b) Compute . −1/x )2
dx.
0 x (1 − e
2

P2022-4 Let A be a square matrix, and suppose positive integers m and n exist so
that .Am = I and .An /= I . Find

det(I + A + A2 + ... + Am−1 ).

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 59

P2022-5

(a) Define .n? as the sum of integers from 1 to n, for example, .5? = 1+2+3+4+5.
Compute the number of zeros that appear at the end of decimal representation
of .2022?.
(b) Define .n! as the product of integers from 1 to n, for example, .5! = 1 ∗ 2 ∗
3 ∗ 4 ∗ 5. Compute the number of zeros that appear at the end of the decimal
representation of .2022!.

P2022-6 Can a group be the union of two of its proper subgroups?

P2022-7 For a function .g : [0, 1] → [0, 1], denote .g ◦ g ◦ ... ◦ g (m times) as .g m .

Suppose that .g : [0, 1] → [0, 1] is continuous and that there is an m so that for all
x, .g m (x) = x. Show that in fact .g 2 (x) = x.

P2022-8 It is well known that .N = {1, 2, 3, . . .} and .N × N have the same

cardinality, and the standard classroom demonstration of this involves a diagonal
lines argument. Explicitly give a function between .N and .N × N, and show that it is
bijective.
The 2022 Problem Solutions begin on page 187 in Chap. 44.
Chapter 23
2023 Problems

P2023-1 A particle is moving along a track. At time t (in seconds) its distance from
one end of the track is .x(t) = 9 + 7t + 3t 2 − t 3 /3 (in feet). What is the fastest speed
that the particle goes during the time interval .0 ≤ t ≤ 10, and when is this? What is
the slowest speed the particle ever goes, and when is this?

P2023-2 Verify that the following equation is true:

sin 50◦ sin 60◦ sin2 70◦

2 cos 70◦ =
. + − .
sin 60◦ sin 50◦ sin 50◦ sin 60◦

P2023-3 For real numbers x, y, and z, define .f (x, y, z) to be the maximum of x,

y, and z, that is, .f (x, y, z) = max{x, y, z}. Evaluate
⎰ 1⎰ 1⎰ 1
. f (x, y, z) dx dy dz.
0 0 0

P2023-4 Find a differentiable function f such that

⎰ x
. f (t) dt = f (x) − f (−x)
−x

and .f (0) = 2023.

P2023-5 Let .f1 , f2 , f3 , . . . be the Fibonacci sequence 1, 1, 2, 3, 5, . . . , in which

fn+1 = fn + fn−1 . Prove that
.

fn+1 fn−1 − fn2 = (−1)n

. for all n ≥ 2.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 61

P2023-6 Consider .ΔABC with its inscribed circle .γ . Assume that the altitude .CH
does not pass through the center of .γ and that H is between A and B. Let X and Y
be the intersections of .CH and .γ as shown. Prove that .CX > H Y .

P2023-7 You have four playing cards. You verify that three of them are red suits
and one is black. You give them to your friend. While you are not watching, your
friend mixes them randomly and places them face down in four locations on a table.
You choose one of the cards. You tell your friend your choice, but you do not look
at the card. Your friend looks at the three cards you did not choose and turns one of
the red cards face up. For each card still face down, what is the probability that it is
red?

P2023-8 There are n small white beads and k black beads moving along an infinite
wire without gravity or friction. The white beads are all to the left of the black
beads. (The figure shows the case .n = 3, .k = 5.) Initially the white beads move to
the right, and the black beads move to the left, all with the same speed. When two
beads collide, they bounce, reversing directions but keeping the same speed.

(a) After the collisions are done, how many beads will be moving to the right and
how many beads will be moving to the left?
(b) How many collisions will there be?

Note: This is not meant to be a physics problem, so please avoid concepts such as
mass, energy, and momentum in your answer.
The 2023 Problem Solutions begin on page 193 in Chap. 45.
Part II
Solutions
Chapter 24
2001 Solutions

The 36th competition was held at the University of Indianapolis.

The 2001 problem set was prepared by Professor Mike Axtell at Wabash College
and Professor Joe Stickles at the University of Evansville.
The 2001 problem statements begin on page 3 in Chap. 1.

S2001-1 The sum of the first a positive integers is given by the sequence of
triangular numbers,

(0, 1, 3, 6, 10, . . . , Ta , . . .),

defined by .Ta = 12 a(a + 1) or recursively by .T0 = 0 and .Ta+1 = Ta + a + 1.

Then the sum of any list of consecutive positive integers is a difference of triangular
numbers: For .0 ≤ a < b,

. (a + 1) + (a + 2) + (a + 3) + · · · + (b − 1) + b. (24.1)
= Tb − Ta
1 1
= b(b + 1) − a(a + 1)
2 2
1
= (b + a + 1)(b − a). (24.2)
2
For a positive integer n which is not a power of 2, there are some numbers .k ≥ 0
and .m ≥ 1 so that .n = 2k (2m + 1). The factors .2k and .2m + 1 can correspond to
the two factors from (24.2), so solving for a and b leads to these two cases:

Case 1. If .m+1 ≤ 2k , then the sum of the positive numbers from .2k −m to .2k +m
is

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 65

. T2k +m − T2k −m−1

1 k
= ((2 + m) + (2k − m − 1) + 1)((2k + m) − (2k − m − 1))
2
1
= (2k+1 )(2m + 1) = n.
2

Case 2. If .m + 1 > 2k , then the sum of the positive numbers from .m + 1 − 2k to

.m + 2 is
k

. T2k +m − Tm−2k
1 k
= ((2 + m) + (m − 2k ) + 1)((2k + m) − (m − 2k ))
2
1
= (2m + 1)(2k+1 ) = n.
2
In either case, there are at least two consecutive numbers in the sum
(24.1); in particular, if n is odd, then .k = 0 and .n = m + (m + 1).
In (24.2), one of the factors must be odd, so if the product is power of 2,
then .b − a = 1 and the sum (24.1) has only one term.

Remark from the Editors This well-known problem appears in [RL], where it is
attributed to a 1976 Canadian Olympiad.

S2001-2 Inflection points of smooth functions are where .f '' (x) changes sign. For
.x > 0, .x 1/x /x 4 > 0, so we must determine the number of times that .g(x) =
((ln(x))2 + (2x − 2) ln(x) + (1 − 3x)) changes sign. Let .h(x) = g(ex ) = x 2 +
(2ex − 2)x + 1 − 3ex for .x ∈ R. .h(x) changes sign when .g(ex ) changes sign. We
see that .h' (x) = (ex + 1)(2x − 1) − 1 and .h'' (x) = ex (2x + 1) + 2 > 0, so .h'
is increasing and has at most one zero. .h' (1/2) = −1 < 0 and .h' (1) = e > 0,
so .h' (x) has exactly one zero. By Rolle’s Theorem, .h(x) has at most two zeros.
.h(−1) = 4 −
e > 0, .h(0) = −2 < 0, and .h(2) = 1 + e > 0. Therefore .h(x)
5 2

changes sign exactly twice. Hence .f '' (x) changes sign exactly twice, and .f (x) has
exactly two inflection points.

S2001-3 Drop one ball into each slot, leaving .k − n balls. Now the question is how
many ways can the remaining .k − n balls be allocated into the n slots. This problem
is in one-to-one correspondence with arrangements of the .k − n balls along with
.n − 1 “dividers.”

To see an example of this, presume .k = 8 and .n = 4, meaning there are four balls
remaining, which we will represent with an o, and three “dividers,” which we will
represent with an x. Note that the arrangement
24 2001 Solutions 67

ooxoxox
.

would correspond with adding two balls to the first slot, one to the second, one to
the third, and zero to the fourth slot. Similarly, the arrangement

xxoooox
.

would correspond with adding zero balls to the first slot, zero to the second, four
to the third, and zero ⎛to the
⎞ fourth slot. In this case, the number of possible
7
arrangements would be . .
3
In general, the number of arrangements is the number of ways one⎛can select
⎞ the
k−1
.n − 1 dividers from the .(k − n) + (n − 1) = k − 1 places, which is . .
n−1

S2001-4A

(a) We apply Euler’s Identity, .eiθ = cos(θ ) + i sin(θ ) to .θ and .3θ to get

. cos(3θ ) + i sin(3θ ) = e3iθ = (eiθ )3

= (cos(θ ) + i sin(θ ))3
= cos3 (θ ) − 3 cos(θ ) sin2 (θ )
⎛ ⎞
+ i 3 cos2 (θ ) sin(θ ) − sin3 (θ ) .

Therefore

. cos(3θ ) = cos3 (θ ) − 3 cos(θ ) sin2 (θ ) (24.3)

= cos (θ ) − 3 cos(θ )(1 − cos (θ ))
3 2

= 4 cos3 (θ ) − 3 cos(θ ).

Thus .f (x) = 4x 3 − 3x.

(b) . 12 = cos(π/3) = f (cos(π/9)). Thus .cos(π/9) is a root of the polynomial
with integer coefficients .p(x) = 2(4x 3 − 3x − 1/2), which can be rewritten
as a composite .g(2x) = (2x)3 − 3(2x) − 1. By the Rational Root Theorem,
any rational root of .g(x) = x 3 − 3x − 1 must be .±1. Since .g(1) = −3 and
.g(−1) = 1, .g(2x) has no rational roots, and .cos(π/9) is irrational.

S2001-4B This alternate approach to part (a) does not use complex numbers but
assumes knowledge of the angle addition formula for cosine and the double angle
formula for sine.
68 24 2001 Solutions

. cos(3θ ) = cos(2θ + θ )
= cos(2θ ) cos(θ ) − sin(2θ ) sin(θ )
= (cos2 (θ ) − sin2 (θ )) cos(θ ) − (2 sin(θ ) cos(θ )) sin(θ )
= cos3 (θ ) − 3 sin2 (θ ) cos(θ ),

which matches (24.3).

Remark from the Editors The statement of the Rational Root Theorem is that if a
degree n polynomial with integer coefficients .an x n + . . . + a1 x + a0 has a rational
root equal to a fraction in the lowest terms .r/s, then r is a divisor of .a0 , and s is a
divisor of .an . A discussion of this theorem appears in the solution to Problem .#1 of
the 1907 Eötvös Competition in Hungary [K2 ] and also in [RL].

S2001-5 Let .Q = (0, 0) be the center of the larger circle .C2 , and without loss of
generality we may rotate the circle so that the path passes through .(2, 0). Let .Ot be
−−→
the point at the center of .C1 at time t, with position vector .QOt parametrized by
.(cos(t), sin(t)) for .0 ≤ t ≤ 2π . At time .t = 0, set .P = (2, 0), so ./ QO0 P = π

radians. In one revolution around Q, the smaller circle rolls along a path twice its
−−→
circumference, so at time t, ./ QOt P = π − 2t as in the figure, and .Ot P is in the
direction .(cos(−t), sin(−t)). Adding vectors,
−→ −−→ −−→
QP = QOt + Ot P = (cos(t), sin(t)) + (cos(−t), sin(−t)) = (2 cos(t), 0),
.

so the point P oscillates on a diameter segment.

Ot 2t
t
Q P

Remark from the Editors This problem is similar to Problem .#3 of the 1926 Eötvös
Competition in Hungary [K2 ].

S2001-6 We observe that for each integer k the interval .[kπ + π/3, kπ + 2π/3] has
length .π/3 > 1 and so must contain at least one integer. The absolute value of that
integer is less than 4k when .k > 1. For the integers inside .[kπ + π/3, kπ + 2π/3]
24 2001 Solutions 69

√
| sin(n)| 3/2
for some k, we see that . > . Hence
n 4k
∞ ∞ √
⎲ | sin(n)| ⎲ 3/8
. ≥ .
n k
n=1 k=2

Since the Harmonic Series diverges to infinity, the comparison test shows that
⎲∞
| sin(n)|
. is infinite.
n
n=1
Chapter 25
2002 Solutions

The 37th competition was held at Anderson University.

The 2002 problem set was prepared by Professor Mike Axtell at Wabash College
and Professor Joe Stickles at the University of Evansville.
The 2002 problem statements begin on page 5 in Chap. 2.

S2002-1

(a) An example of such a function would be .f (x) = x − b2 .

(b) Note that the assumption that f is positive and continuous on .[0, b] implies that
the expression is well defined and integrable.
⎰ ⎰
b f (x) f (x) + f (b − x) − f (b − x)
b
. dx = dx
0 f (x) + f (b − x) 0 f (x) + f (b − x)
⎰ b⎛ ⎞
f (b − x)
= 1− dx
0 f (x) + f (b − x)
⎰ b
f (b − x)
=b− dx.
0 f (x) + f (b − x)

Now, let .u = b − x. Then, .du = − dx. Also, when .x = 0, we have .u = b, and

when .x = b, we have .u = 0. Hence, we have
⎰ b ⎰ 0
f (x) f (u)
. dx = b − (−du)
0 f (x) + f (b − x) b f (b − u) + f (u)
⎰ b ⎰ b
f (x) f (u)
=⇒ dx = b − du
0 f (x) + f (b − x) 0 f (u) + f (b − u)

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 71

⎰ b ⎰ b
f (x) f (u)
=⇒ dx + du = b.
0 f (x) + f (b − x) 0 f (u) + f (b − u)

The variables make no difference in the integrals, so the left-hand side becomes
⎰ b f (x)
.2 dx = b
0 f (x) + f (b − x)
⎰ b f (x) b
=⇒ dx = .
0 f (x) + f (b − x) 2

Remark from the Editors This problem and solution are similar to Problem .#5 in
[HWi].

S2002-2 Since each .ai > 0 for all .i ≥ 1, we have

a1 + a2 + · · · + an a1 1
bn =
. ≥ = a1 · .
n n n
∑ ∑
Since the series . n1 diverges, the series . an1 diverges (using .a1 /= 0), and hence
∑
. bn diverges by the direct comparison test.

S2002-3 Let p and q be consecutive odd primes with .p < q. Then, .p + q is even
and can be factored into these two positive integers:
⎛ ⎞
p+q
.p+q =2 .
2

Now, the fraction . p+q

2 represents the average of p and q, and since .p < q, we have
p+q
.p <
2 < q. Since these two primes were consecutive, we must have . p+q 2 is
composite, and hence it has at least two (possibly repeated) prime factors. Thus, the
sum of two consecutive odd primes has at least three prime factors. These factors
may repeat, considering the examples .3 + 5 = 23 , .5 + 7 = 22 · 3, and .7 + 11 = 2 · 32 ,
or there may be exactly three prime factors, .13 + 17 = 2 · 3 · 5.

Remark from the Editors This problem appears, with a similar solution, as Problem
#80 in [HWi].
.

S2002-4 Let .an denote the total number of books sold from day 1 through day n.
Hence, we have

1 ≤ a1 < a2 < a3 < · · · < a100 = 140.

.
25 2002 Solutions 73

Adding 59 to each above term gives

60 ≤ a1 + 59 < a2 + 59 < a3 + 59 < · · · < a100 + 59 = 199.

Now, the list .a1 , a2 , ..., a100 , a1 + 59, a2 + 59, ..., a100 + 59 is a list of 200 numbers
that lie between 1 and 199, inclusive. By the pigeonhole principle (also known as
Dirichlet’s box principle), there must be two numbers in the list that have the exact
same value. Since .a1 < a2 < · · · < a100 , we know .ai /= aj for any .i /= j . Similarly,
we cannot have .ai + 59 = aj + 59 for .i /= j . Hence, we must have .ai = aj + 59
for some i, j , with .i > j by the increasing property of the list, and .ai − aj = 59.
Because .ai − aj represents the number of books sold from day .j + 1 to day i, there
were exactly 59 books sold from day .j + 1 until day i.

Remark from the Editors This problem and solution are similar to Problem .#22 in
[HWi].

S2002-5 In some steps, multiplication is denoted with the .· (dot).

(a) Let e be the identity of S. If there exists an .x ∈ S such that .ϕ (x) = 0, then
⎛ ⎞ ⎛ ⎞ ⎛ ⎞
ϕ (e) = ϕ xx −1 = ϕ (x) ϕ x −1 = 0 · ϕ x −1 = 0,
.

and for any other .y ∈ S, we have

ϕ (y) = ϕ (ey) = ϕ (e) ϕ (y) = 0 · ϕ (y) = 0,

and hence .ϕ (x) = 0 for all .x ∈ S.

Now, suppose .ϕ (x) /= 0 for all .x ∈ S. Then,

ϕ (e) = ϕ (ee) = ϕ (e) ϕ (e)

=⇒ 0 = ϕ (e) ϕ (e) − ϕ (e) = (ϕ (e) − 1)ϕ (e)

=⇒ ϕ (e) = 1 (since ϕ (e) /= 0),

and for any .x ∈ S, we have

⎛ ⎞
ϕ (e) = ϕ xx −1
.

⎛ ⎞
=⇒ 1 = ϕ (x) ϕ x −1 .

( )
Since .ϕ (x) and .ϕ x −1 are nonnegative integers with product 1, we must have
.ϕ (x) = 1. Hence, .ϕ (x) = 1 for all .x ∈ S.

(b) If .ϕ (0S ) /= 0, then for all .x ∈ S, we get

ϕ (0S ) = ϕ (0S · x) = ϕ (0S ) ϕ (x)

.
74 25 2002 Solutions

=⇒ ϕ (x) = 1,

which contradicts .ϕ being nonconstant. Hence, we must have .ϕ (0S ) = 0.

S2002-6 First, for the four corners of an inscribed square to lie on the three sides
of the triangle, two corners must lie on the same side, and they cannot be opposite
corners, so (at least) one square side is a subinterval of one triangle side as in the
figure. Next, we need the fact that for any numbers b and h, we have

. (b − h)2 ≥ 0
b2 − 2bh + h2 ≥ 0
b2 + h2 ≥ 2bh.

Now, we notice that the square “divides” the triangle into a trapezoid with bases b
and s and a height of s and a triangle with base s and height .h − s. We must have
the sum of these areas equal the total area of the triangle; hence, we have

1 1 1
. s (b + s) + s (h − s) = bh
2 2 2
bs + s 2 + sh − s 2 = bh
s (b + h) = bh
s (b + h)2 = b2 h2
2
⎛ ⎞
s 2 b2 + 2bh + h2 = b2 h2

s 2 (2bh + 2bh) ≤ b2 h2 by above

s 2 (4bh) ≤ b2 h2
⎛ ⎞
1 1 1
s2 ≤ bh = bh .
4 2 2

Hence, the area of the square is at most one half the area of the triangle.

h
s

b
Chapter 26
2003 Solutions

The 38th competition was held at Butler University.

The 2003 problem set was prepared by Professor Mike Axtell at Wabash College
and Professor Joe Stickles at the University of Evansville.
The 2003 problem statements begin on page 7 in Chap. 3.

The 2003 competition’s first place team from Indiana—Purdue Fort Wayne. Left to Right: Jeff
Wilkins, Christian MacLeod, Coach Doug Weakley, and Kevin Chlebik. Photo courtesy of Purdue
Fort Wayne Communications and Marketing [PFW]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 75

S2003-1

(a) Since .(a ∗ b) ∗ (a ∗ b) = a ∗ b by the given property, we have by substitution

that .c ∗ c = c.
(b) We have .(a ∗ b) ∗ (c ∗ y) = a ∗ y by the given property. Using the hypothesis,
we get .c ∗ (c ∗ y) = a ∗ y. From part (a), we can substitute .c ∗ c in for c, yielding
.(c ∗ c) ∗ (c ∗ y) = a ∗ y. Finally, using the given property, we get .c ∗ y = a ∗ y.

S2003-2 Let .pk represent the probability a particular pair of competitors will face
each other in a tournament involving .2k competitors. By induction on k, .k = 1:
1 1
There are only .21 competitors, so .p = 1 = k−1 . Assume .pk = k−1 for .k ≤ n.
2 2
For .k = n + 1, the probability a competitor meets any other competitor in the first
1
round is . k+1 . The probability they meet in later rounds is the probability they
2 −1
do not meet in round 1,⎛they both win⎞in⎛round ⎞ ⎛ 1,⎞ and they meet in subsequent
1 1 1
rounds. This is given by . 1 − k+1 pn . So,
2 −1 2 2
⎛ ⎞⎛ ⎞⎛ ⎞
1 1 1 1
pn+1 =
. + 1− pn
2n+1 − 1 2k+1 −1 2 2
2n+1 − 2 1
1 1
= + · · n−1 by ind. hyp.
2n+1 −1 2 n+1 −1 4 2
1 2n − 1
= n+1 +
2 − 1 2n (2n+1 − 1)
2n + 2n − 1
=
2n (2n+1 − 1)
1
= .
2n
1
By induction, .pk = for all .k ≥ 1.
2k−1
S2003-3A If we connect the five points on the unit circle, we get a regular
pentagon. Hence, the length a is the same as the one shown here.
26 2003 Solutions 77

1
a 72°

√
From the Law of Cosines, .a = 2 − 2 cos 72◦ . Going back to the given chords, we
get the following picture:

36°
a
108°
b
a

Again using the Law of Cosines, we get .b2 = 2a 2 − 2a 2 cos 108◦ = 4 − 4 cos 72◦ −
(4−4 cos 72◦ ) cos 108◦ = 4−4 cos 72◦ +(4−4 cos 72◦ ) cos 72◦ = 4−4 cos2 72◦ =
4 sin2 72◦ so .b = 2 sin 72◦ . Using the Law of Sines, we get

sin 108◦ sin 36◦

. =
b a
sin 72◦ sin 36◦
◦
=
2 sin 72 a
a = 2 sin 36◦ .

So,

.ab = (2 sin 36◦ )(2 sin 72◦ ) = 4(2 sin 18◦ cos 18◦ )(cos 18◦ )
= 8 sin 18◦ cos2 18◦ (26.1)
◦ ◦
= 8 sin 18 (1 − sin 18 ). 2
78 26 2003 Solutions

S2003-3B In the first figure above, by drawing the diameter of the circle that bisects
the segment of length a, one can find a right triangle with an .18◦ angle opposite
to side length . 12 a and with hypotenuse b, so .sin 18◦ = 2b a
. In the second figure,
drawing the radius that bisects the segment of length b results in a right triangle
with .18◦ angle adjacent to side length . 12 b and with hypotenuse 1, so .cos 18◦ = b2 .
Then .ab = (2b sin 18◦ )(2 cos 18◦ ), leading to the above line (26.1).

S2003-4 Write the square as an array:

a b c
. d e f
g h i

k k k
Then .k = aei = beh = ceg. So, .a = , .b = , .c = . Now,
ei eh eg

k k k k3 k2
k = abc =
. · · = 3 = 3,
ei eh eig e ghi e

and solving for k gives .k = e3 .

Remark from the Editors See S2017-4 in Chap. 40 for an alternate solution.

S2003-5 Considering some partial sums,

⎛ ⎞
1 1
S3 = 1 +
. − ≥1
2 3
⎛ ⎞ ⎛ ⎞
1 1 1 1 1 1
S6 = 1 + − + + − ≥1+
2 3 4 5 6 4
⎛ ⎞ ⎛ ⎞ ⎛ ⎞
1 1 1 1 1 1 1 1 1 1
S9 = 1 + − + + − + + − ≥1+ + .
2 3 4 5 6 7 8 9 4 7

1 1 1 ∑
The pattern is .S3n ≥ 1 + + + · · · + . A limit comparison test with . n1
∑ 1 4 7 3n − 2
shows . 3n−2 diverges. Hence, a direct comparison shows the given series has
unbounded partial sums and diverges.
26 2003 Solutions 79

S2003-6
⎾ ⏋ ⎾ ⏋
21 f3 f2
(a) By induction on n, for .n = 2, .Q
2= = . Assume .Qn =
11 f2 f1
⎾ ⏋ ⎾ ⏋⎾ ⏋ ⎾ ⏋
fn+1 fn fn+1 fn 11 fn+2 fn+1
. For .n+1, .Q n+1 =Q Q=
n = .
fn fn−1 fn fn−1 1 0 fn+1 fn
So, the result holds. ⎾ ⏋
f3n+1 f3n
(b) From part (a), .Q =
3n . Also,
f3n f3n−1

⎾ ⏋3
fn+1 fn
.Q3n = Qn Qn Qn = .
fn fn−1

Direct computation shows that the upper right entry when cubing this matrix
2 + f 2 )f + f
n+1 fn fn−1 + fn fn−1 . Equating this with the upper right
is .(fn+1 2
n n
entry of the first matrix, we get

.f3n = (fn+1
2
+ fn2 )fn + fn+1 fn fn−1 + fn fn−1
2

= fn+1
2
fn + fn3 + fn+1 fn fn−1 + fn fn−1
2

= fn+1
2
(fn+1 − fn−1 ) + fn3
+fn+1 (fn+1 − fn−1 )fn−1 + (fn+1 − fn−1 )fn−1
2

= fn+1
3
− fn+1
2
fn−1 + fn3 + fn+1
2
fn−1 − fn+1 fn−1
2
+ fn+1 fn−1
2
− fn−1
3

= fn+1
3
+ fn3 − fn−1
3
.
Chapter 27
2004 Solutions

The 39th competition was held at Indiana State University.

The 2004 problem set was prepared by Professor Mike Axtell at Wabash College
and Professor Joe Stickles at the University of Evansville.
The 2004 problem statements begin on page 9 in Chap. 4.

S2004-1 We attempt to partition .{1, 2, 3, 4, 5} into two sets A and B in such a way
that neither set contains two numbers and their difference. Thus, 2 cannot be in the
same set as either 1 or 4; else we would have .2 − 1 = 1 or .4 − 2 = 2. So, put 2 in
A, and put 1 and 4 in B. If we put 3 in B, then we have .4 − 3 = 1. So, 3 must go in
A. Similarly, placing 5 in B leads to .5 − 4 = 1; thus, 5 cannot be in B. However, 5
cannot be in A since .5 − 3 = 2. We have reached a contradiction. Hence, no matter
how the two sets are constructed, one of the two sets must contain two numbers and
their difference.

Remark from the Editors This problem and solution appear as Problem .#3 of the
1916 Eötvös Competition in Hungary [K2 ].

S2004-2

(a) Since
|⎛ ⎛ ⎞
| 2n+1 ⎞ ⎛ 2 n ⎞|
| 2n+1 a 2n + 1
. lim | / | = lim ·
n→∞ | a 2n+1 + 1 a 2 + 1 | n→∞ 2n a 2n+1 + 1
n

⎾ ⏋
1 2n 1
= 2 lim (1 + 2n )/(a + 2n )
n→∞ a a
= 2(0) = 0,

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 81

this series converges by the Ratio Test.

(b) Since
n
2n 2n (a 2 − 1)
. =
a2 + 1 (a 2 + 1)(a 2 − 1)
n n n

n
2n (a 2 − 1) + 2n+1 − 2n+1
=
(a 2 + 1)(a 2 − 1)
n n

n
2n (a 2 + 1) − 2n+1
=
(a 2 + 1)(a 2 − 1)
n n

2n 2n+1
= − n+1 ,
− 1 a2 − 1
n
a2
∞
⎲ ⎲∞ ⎛ ⎞
2n 2n 2n+1
the sum . = − is telescoping (with the
a2 + 1 a 2 − 1 a 2n+1 − 1
n n
n=0 n=0
last terms in partial sums approaching 0 by part (a)). Hence
∞
⎲ 2n 20 1
. = 0 = .
a +1
2n
a −1
2 a − 1
n=0

S2004-3 Let B be the matrix obtained from A by subtracting row 1 of A from each
of the other three rows. Then .det A = det B. Each entry in the last three rows of B
is .−3, 0, or 3 and therefore divisible by 3. Now let C be the matrix obtained from B
by dividing each of the entries in the last three rows of B by 3. All of the entries of
C are integers, giving .det C is an integer; moreover, .det A = det B = 33 det C. So,
.det A is divisible by 27.

Remark from the Editors This problem and solution appear as Problem .#10 in
[GKL].

S2004-4 Induction on n, the length of the sequence, starts with .n = 1, where the
sequence is

1−1 ⎛
⎲ ⎞
1−1
ai+1
i
i=0
a1 =
. .
21−1
So, assume that for any sequence of length k, with .k ≤ n, we will have final term
27 2004 Solutions 83

k−1 ⎛
⎲ ⎞
k−1
ai+1
i
i=0
. .
2k−1
For .k = n + 1, after the first step, we have the sequence

a1 + a2 a2 + a3 an + an+1
. , ,..., ,
2 2 2
which is a sequence of length n. Thus, by the inductive hypothesis, the final term
will be
n−1 ⎛
⎲ ⎞⎛ ⎞
n−1 ai+1 + ai+2
i 2
i=0
.
2n−1
⎛n−1 ⎛ ⎞ ⎛n−1 ⎛ ⎞
⎲ n − 1⎞ ⎲ n − 1⎞
ai+1 + ai+2
i i
i=0 i=0
=
2n
⎛n−1 ⎛ ⎞ ⎛n−2 ⎛ ⎞
⎲ n − 1⎞ ⎲ n − 1⎞
a1 + an+1 + ai+1 + ai+2
i i
i=1 i=0
=
2n
⎛n−2 ⎛ ⎞ ⎛n−2 ⎛ ⎞
⎲ n − 1⎞ ⎲ n − 1⎞
a1 + an+1 + ai+2 + ai+2
i+1 i
i=0 i=0
=
2n
n−2 ⎛⎛
⎲ ⎞ ⎛ ⎞⎞
n−1 n−1
a1 + an+1 + + ai+2
i+1 i
i=0
=
2n
n−2 ⎛
⎲ ⎞
n
a1 + an+1 + ai+2
i+1
i=0
=
2n
n−1 ⎛ ⎞
⎲ n
a1 + an+1 + ai+1
i
i=1
=
2n
n ⎛ ⎞
⎲ n
ai+1
i
i=0
= .
2n
84 27 2004 Solutions

S2004-5 Since P is the center of the square, .ΔAP C will also be a right triangle.
Construct a circle with diameter .AC; both points B and P will be on this circle.
(This is because the circle that circumscribes a right triangle has as its center the
midpoint of the hypotenuse.) Since .AP and .P C are equal chords of this circle, the
arcs AP and P C are equal. Thus ./ ABP = / CBP .

B
A C

S2004-6 Since both boats remain in their slips for the same amount of time, this
information does not enter into the solution of the problem. When the ferryboats
meet for the first time, the combined distance the boats have traveled is equal to the
width of the river. When the boats reach the opposite shore, the combined distance
the boats have traveled equals two widths of the river. When they then meet the
second time, the combined distances the boats have traveled are three widths of the
river. Since the boats move at a constant speed, it follows that each boat has traveled
three times as far as when they first met and had traveled a combined distance of
one river width. The slow boat had traveled 720 yards when the boats first met.
Thus, by the second meeting, the slow boat has traveled .3 × 720 = 2160 yards.
Since this second meeting occurs at the point when the slow boat has moved 400
yards away from the far shore, it follows that the width of the river is given by
.2160 − 400 = 1760 yards.
Chapter 28
2005 Solutions

The 40th competition was held at Indiana University—Purdue University Fort

Wayne for the first time.
The 2005 problem set was prepared by Professor Dan Coroian at the host institution.
The 2005 problem statements begin on page 11 in Chap. 5.

Professor David Housman, Goshen College, presents a book prize to the 2005 first place team from
Purdue University

S2005-1 Since x is in .[−1, 1], we can make the substitution .θ = cos−1 x, with
.θ ∈ [0, π ]. The integral becomes

⎰ ⎰ ⎰
1 | sin(n cos−1 x)| 0 π
. √ dx = − | sin nθ | dθ = | sin nθ | dθ.
−1 1 − x2 π 0

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 85

/3 2 /3

From the graph of .| sin nθ | (seen above for .n = 3), the integral can be calculated as
⎰ ⎰ |
π π/n cos nθ ||π/n
| sin nθ | dθ = n sin nθ dθ = −n = 2.
n |0
.
0 0

S2005-2 Toward a contradiction, assume that there exists a red–white–blue color-

ing of the points in the plane such that no two points of the same color are at distance
1 from each other. We will show √ that for such a coloring there are no two points of
different colors at a distance of . 3 units. Indeed, suppose such a pair does exist.
Without loss of generality, we can assume it is a red–white pair and denote the two
points R and W , respectively. If we construct a rhombus RBW A with side lengths
1 and diagonal RW , we see that the other two vertices, A and B, must be colored
blue. However, it is easy to see that diagonal AB has length 1, which contradicts our
original
√ assumption. Therefore, for the coloring in question, all points at a distance
. 3 have the same color. This implies that the circle .x + y
2 2 = 3 will have to

be colored the same color as the origin, which leads to a contradiction with our
assumption that no two points of the same color are at distance 1 from each other
(just choose two points on this circle that are distance 1 from each other).

S2005-3 We have

1 ⎲ ∏
n n
1
0 ≤ an =
. ln(1 + kn) = ln (1 + kn)
n3 n3
k=1 k=1

1 ∏
n
1 1
≤ ln (1 + n2 ) = ln(1 + n2 )n = 2 ln(1 + n2 )
n3 n3 n
k=1
1 2 2 2n + 2
≤ ln(1 + n)2 = 2 ln(1 + n) ≤ 2 (1 + n) = .
n2 n n n2
2n + 2
Thus, .0 ≤ an ≤ , so using the Squeeze Theorem, . lim an = 0.
n2 n→∞
28 2005 Solutions 87

S2005-4A If we let .b = a and .d = 1/a in the first equation, we obtain .(a ∗ a)(c ∗
(1/a)) = (ac) ∗ 1 = ac, that is, .c ∗ (1/a) = ca. If we now use .1/a instead of a,
we obtain .c ∗ a = c/a, so operation .∗ is just the usual division of rational numbers.
Therefore, .((6/5) ∗ (8/15)) ∗ 2 = ((6/5) ÷ (8/15)) ÷ 2 = (6/5)(15/8)(1/2) = 9/8.

S2005-4B Lemma: .x ∗ y = x ÷ y. Proof: Step 1. .y · (1 ∗ y) = (y ∗ 1) · (1 ∗ y) =

(y · 1) ∗ (1 · y) = y ∗ y = 1, so .1 ∗ y is the reciprocal of y (by uniqueness of inverse
in .Q+ ). Step 2. .x ∗ y = (x · 1) ∗ (1 · y) = (x ∗ 1) · (1 ∗ y) = x · y −1 .
From the lemma,

((6/5) ∗ (8/15)) ∗ 2 = ((6/5) ÷ (8/15)) ÷ 2 = (6/5)(15/8)(1/2) = 9/8.

Remark from the Editors In both of the above solutions, showing that the operation
is the same as .÷ also shows the existence of such an operation .∗, so that the three
given properties are not logically inconsistent.

S2005-5A Since .B = I −A, .AB +I = A(I −A)+I = I +A−A2 , and .I −A+A2

is an inverse:

. (I + A − A2 )(I − A + A2 )
= (I − A + A2 ) + A(I − A + A2 ) − A2 (I − A + A2 )
= I − A + A2 + A − A2 + A3 − A2 + A3 − A4
= I − A2 + A3 + A3 − A4 .

The last four terms cancel because .A4 = A3 A = A2 A = A3 = A2 . One should

also cite the linear algebra fact that in the ring .Mn (R) it is enough to check just one
product or explicitly check that the product in the other order also gives the identity
by an analogous computation (then, this solution works for any ring).

S2005-5B Multiply .A+B = I by A on each side, and use the assumption .A2 = A3 ,
to see that .ABA = 0. Then .I − AB is an inverse:

(AB + I )(I − AB) = AB − (AB)(AB) + I 2 − AB = I − (ABA)B = I

(and similarly for the product in the other order).

S2005-6 We have
√
. 9x 4 − 24x 3 + 6x 2 + 5 − (ax + bx + c)
p

⎾/ ⎛ ⎞⏋
24 6 5 b c
=x 2
9− + 2 + 4 −x p−2
a + p−1 + p .
x x x x x
88 28 2005 Solutions

If .p /= 2 or if .p = 2 and .a /= 3, then the limit is .±∞, so .p = 2 and .a = 3.

Substituting these values and rationalizing the numerator the limit become

9x 4 − 24x 3 + 6x 2 + 5 − (3x 2 + bx + c)2

. lim √
x→∞ 9x 4 − 24x 3 + 6x 2 + 5 + (3x 2 + bx + c)

(−24 − 6b)x 3 + (6 − b2 − 6c)x 2 + · · ·

= lim ⎛/ ⎞.
x→∞ 2
x 9 − 24
x + x2 + x4 + 3 + x + x2
6 5 b c

For the limit to be equal to .7/3, we must have .−24 − 6b = 0 and . 6−b6−6c =
2 7
3,
which gives .b = −4 and .c = −4.

S2005-7 Extend AM until it crosses DC in .D ' , and extend BN until it crosses AD

in .A' . Triangle ABP is similar to .D ' N P (since .D ' N is parallel to AB). Then

BP AB AB AB 2
. = = = = ,
NP N D' N C + CD ' 1
2 AB + AB 3

since .CD ' = AB, because of the fact that triangles .MCD ' and MBA are congruent
+BP = 3+2 , i.e., . BN = 5 . In the same
2 2
(A.S.A.). But from this equation we get . N PBP BP

way, since triangles BP M and .P AA' are similar, we have

1
MP BP MB MB AD 1
. = '
= '
= = 2 = .
AP PA AA 2AD 2AD 4

(Here we used the fact that .AD = DA' , since triangles .A' DN and BN C are
+AP
congruent.) Therefore . MPAP = 1+4
4 , i.e., . AM = 5 .
AP 4

S2005-8 We write .x = 5k + x1 , .y = 5l + y1 , and .z = 5m + z1 , where

.k, l, m, x1 , y1 , z1 are integers, and without loss of generality we can assume that
.0 ≤ x1 ≤ y1 ≤ z1 < 5. If 25 divides .x + y + z , then using the Binomial Theorem
5 5 5

we see that 25 must also divide .x1 +y1 +z1 . We will show that .x1 = 0, which implies
5 5 5

25 divides .y 5 + z5 . It follows from expanding the polynomial .(x1 + y1 + z1 )5 that

since 5 divides .x15 + y15 + z15 , 5 also divides .(x1 + y1 + z1 )5 . For any integer a,
we have .a 5 ≡ a (mod 5) by Fermat’s Little Theorem (see [HWr] or [M]), and
so 5 must divide .x1 + y1 + z1 . Toward a contradiction, suppose .x1 > 0. Then
.3 ≤ x1 + y1 + z1 ≤ 12, and there are only four choices for the triple .(x1 , y1 , z1 )

such that 5 divides .x1 + y1 + z1 , namely

(1, 1, 3), (1, 2, 2), (2, 4, 4), and (3, 3, 4).

We have

15 ≡ 1 (mod 25),
.
28 2005 Solutions 89

25 ≡ 7 (mod 25),
35 ≡ −7 (mod 25),
45 ≡ −1 (mod 25).

By inspection, for none of the four choices above is the sum .x15 +y15 +z15 divisible by
25, which represents a contradiction. Therefore, .x1 = 0, and the conclusion follows.
Chapter 29
2006 Solutions

The 41st competition was held at Taylor University for the first time.
The 2006 problem set was prepared by Professor Mike Axtell at Wabash College
and Professor Joe Stickles at the University of Evansville.
The 2006 problem statements begin on page 13 in Chap. 6.

S2006-1 Rewriting, we notice that

√ √
f (x) =
. 16 − (x − 2)2 − 1 − (x − 5)2 .

So, the values of .f (x) are the differences of the y coordinates of two semicircles,
and these semicircles are internally tangent at the point .(6, 0). Furthermore, we
notice that the domain for .f (x) is .[4, 6]. Since the function
√
g(x) =
. 16 − (x − 2)2
√
is decreasing on .[4, 6], the expression
√ . 4x − x + 12 achieves
2
√ its maximum value
at .x = 4, and at this point
√ .g(4) = 4(4) − 42 + 12 = 2 3. Furthermore, at .x = 4,

the function .h(x) = 1 − (x − 5)2 is as small as it can get; √ namely, .h(4) = 0.

Hence the maximum value of .f (x) = g(x) − h(x) is .f (4) = 2 3.

S2006-2 Using the hint, we see for .n ≥ 1,

⎛ ⎞
1
n − 1
n+2
. arctan(1/n) − arctan(1/(n + 2)) = arctan
1+ 1
n(n+2)

= arctan(2/(n + 1)2 ).

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 91

Let .SN denote the N th partial sum of the given series. For .N ≥ 3,

⎲
N ⎛ ⎞ ⎲
N ⎛ ⎞
2 2
SN =
. arctan = arctan 2 + arctan
n2 n2
n=1 n=2

⎲
N −1 ⎛ ⎞
2
= arctan 2 + arctan
(n + 1)2
n=1

⎲
N −1 ⎛ ⎛ ⎞ ⎛ ⎞⎞
1 1
= arctan 2 + arctan − arctan
n n+2
n=1
⎛ ⎞ ⎛ ⎞ ⎛ ⎞
1 1 1
= arctan 2 + arctan 1 + arctan − arctan − arctan .
2 N N +1

As .N → ∞, we have .SN → arctan 2 + arctan 1 + arctan( 12 ) = 3π 4 . The last equality

holds because .arctan 1 = π4 , and for .x > 0, we have .arctan(x) + arctan( x1 ) = π2 .
⎾√ ⏋ √
S2006-3 In order for . 10X = 20, we need .20 ≤ 10X < 21 or .40 ≤ X <
⎾√ ⏋ √
44.1. Given . 3 X = 3, we know .3 ≤ 3 X < 4 or .27 ≤ X < 64. So, we find

⎛⎾√ ⏋ |⎾ √ ⏋ ⎞ ⎛⎰ 44.1 ⎞ ⎛⎰ 64 ⎞
| 3 1 1
.P 10X = 20 | X = 3 = dx / dx
40 x2 27 x
2
⎛ ⎞ ⎛ ⎞
1 1 1 1 984
= − / − = .
40 44.1 27 64 9065

S2006-4 Modeling the position x by the continuous, increasing function .f (t), for
time .0 ≤ t ≤ 72, with .f (0) = 0 and .f (72) = 18, the inverse function .t = f −1 (x)
is continuous and increasing for .0 ≤ x ≤ 18 (the continuity of the inverse is a
well-known fact from calculus which is not proved here). Let .T (x) be the time in
minutes that it takes the rider to go from point x to point .x + 3. Certainly, .T (x) =
f −1 (x + 3) − f −1 (x) is continuous on .[0, 15]. Notice that

T (0) + T (3) + T (6) + T (9) + T (12) + T (15) = 72.

It is not possible for all of the values on the left-hand side to be strictly greater than
12, nor is it possible for all of the values on the left-hand side of the equation to be
strictly less than 12. Hence, there exist .r, s ∈ {0, 3, 6, 9, 12, 15} with .T (r) ≤ 12 and
.T (s) ≥ 12. Since .T (x) is continuous, the Intermediate Value Theorem guarantees

that there exists a k between r and s (or .k = r or .k = s) so that .T (k) = 12. In other
words, there exists a point k so that the rider travels from point k to point .k + 3 in
exactly 12 minutes.
29 2006 Solutions 93

Remark from the Editors This solution uses, in an essential way [W], the fact that
18 miles is an integer multiple of 3 miles.

S2006-5 Toward a contradiction, assume that neither T nor U is closed under

multiplication. Thus, there exist .t1 , t2 ∈ T such that .t1 t2 /∈ T and .u1 , u2 ∈ U
such that .u1 u2 /∈ U . Since .S = T ∪ U , we have .t1 t2 ∈ U and .u1 u2 ∈ T .
Consider the element .t1 t2 u1 u2 . Since the product of any three elements of T
(respectively, U ) is in T (respectively U ); we have .t1 t2 u1 u2 = (t1 t2 )u1 u2 ∈ U
and .t1 t2 u1 u2 = t1 t2 (u1 u2 ) ∈ T . So, .t1 t2 u1 u2 ∈ T ∩ U . However, T and U are
disjoint, a contradiction. Thus, either T or U is closed under multiplication.

S2006-6 Suppose x and y are integers satisfying the given equation. Factoring the
left-hand side of the equation, we see

((r − s)x − (r + s)y)((r + s)x + (r − s)y) = 1.

Since .r, s, x, y are integers, we have the product of two integers, and this product
must be 1. Thus, the factors in the product are either both 1 or both .−1. Solving the
system .(r − s)x − (r + s)y = δ, .(r + s)x + (r − s)y = δ, where .δ = ±1, we obtain
.x =
rδ
r 2 +s 2
and .y = r −sδ
2 +s 2 . For nonzero integers r and s, these solutions satisfy
.0 < |x| < 1 and .0 < |y| < 1, so neither x nor y is an integer, a contradiction.

S2006-7

P
Q

B C

Let .x = P A = P A' and .y = QA = QA' . Then .P B = 3 − x and .QC = 3 − y.

Applying the Law of Cosines to .ΔP BA' , we have .x 2 = 12 + (3 − x)2 − 2(1)(3 −
x) cos 60◦ . Solving for x, we get .x = 75 . Applying the Law of Cosines to .ΔQCA' ,
we have .y 2 = 22 + (3 − y)2 − 2(2)(3 − y) cos 60◦ . Solving for y, we get .y = 74 .
Finally, applying the Law of Cosines to .ΔP A' Q, we have .P Q2 = ( 75 )2 + ( 74 )2 −
√
2( 75 )( 47 ) cos 60◦ . Solving for P Q, we get .P Q = 7 21
20 .
94 29 2006 Solutions

S2006-8 Label the twelve stones .{A, B, C, D, E, F, G, H, I, J, K, L}. Weighing

1 is .{A, B, C, D} versus .{E, F, G, H }. If the scales balance, go to 1. If the scales
do not balance, go to 2.

1. The scales balanced, meaning the odd stone is in .{I, J, K, L}. Weighing 2 is
.{I, J, K} versus the good stones .{A, B, C}. If the scales balance, go to 1a. If

the scales do not balance, go to 1b.

1a. The odd stone must be L. Weighing L against any other stone will determine
whether L is heavier or lighter than the others.
1b. The odd stone is in .{I, J, K}, and it is light (respectively heavy) because
.{I, J, K} is lighter (respectively heavier) than .{A, B, C}. Weighing 3 is I versus

J . If they balance, then K is the odd stone and light (respectively heavy). If they
do not balance, then the light (respectively heavy) side is the odd stone.
2. The scales tipped. Without loss of generality, assume that .{A, B, C, D} was
the heavier side. So, we either have an odd, heavy stone among .{A, B, C, D}
or an odd, light stone among .{E, F, G, H }. Weighing 2 is .{A, B, E} versus
.{F, C, D}. If they balance, go to 2a. If .{A, B, E} is heavier than .{F, C, D}, go

to 2b. If .{A, B, E} is lighter than .{F, C, D}, go to 2c.

2a. The odd stone must be in .{G, H }, and it must be light since .{G, H } were on the
light side of the first measurement. Measure G against H ; the light side is the
odd, light stone.
2b. .{A, B, E} was heavier than .{F, C, D}. Neither C nor D can be an odd, heavy
stone, and E cannot be an odd, light stone. Thus, either A or B is the odd, heavy
stone, or F is the odd, light stone. Weighing 3 is .{A, F } against any two normal
stones (e.g., .{G, H } ). If .{A, F } is heavier, then A is the odd, heavy stone. If
.{A, F } is lighter, then .{F } is the odd, light stone. If they balance, then .{B} is

the odd, heavy stone.

2c. .{A, B, E} was lighter than .{F, C, D}. Neither A nor B can be an odd, heavy,
while F cannot be odd and light. Thus, either C or D is the odd, heavy stone, or
E is the odd, light stone. Weighing 3 is .{C, E} against any two normal stones
(e.g., .{G, H }). If .{C, E} is heavy, then C is the odd, heavy stone. If .{C, E} is
light, then E is the odd, light stone. If they balance, then D is the odd, heavy
stone.

Remark from the Editors Coin balancing problems like this are well-known; for a
survey paper, we refer to [GN]. This version with 12 coins is due to H. Grossman
[G].
Chapter 30
2007 Solutions

The 42nd competition was held at the University of Indianapolis.

The 2007 problem set was prepared by Professor Mike Axtell at Wabash College
and Professor Joe Stickles at Millikin University.
The 2007 problem statements begin on page 15 in Chap. 7.
√ √ √ √
S2007-1 If .x = p + q, then .x − p = q. Squaring both sides, we obtain
√ √
.x − 2x p + p = q, or .−2x p = q − p − x 2 = −x 2 + (q − p). Square both
2

sides again, we get .4px 2 = x 4 − 2(q − p)x 2 + (q − p)2 , so .f (x) = x 4 − 2(q +

p)x 2 + (q − p)2 will suffice.

S2007-2 We know that .gcd(36, n) = x and .lcm(36, n) = 500 + x. Now, since x is

a divisor of 36, we have the following possibilities for x: 1, 2, 3, 4, 6, 9, 12, 18, 36.
Now, 36 must also divide .500 + x, and by experimentation we get .x = 4. This tells
us that .n = 4b and that 4b must divide 504. Hence, b must divide 126. However,
.n = 4b cannot have 3 as a factor, and since .126 = 2 · 3 · 7, the possible values of b
2

are 1, 2, 7, and 14. We quickly see 14 is the only possibility. Hence .n = 4·2·7 = 56.

1 1
S2007-3 For .x ≤ t < 3x, we have .0 < √ ≤ √ . Hence
t t +1
4 x x4 + 1
⎰ 3x ⎰ 3x
dt dt
0<
. √ ≤ √
x t t4 + 1 x x x4 + 1
⎰ 3x
1 1 2
= √ dt = √ · (2x) = √ .
x x4 + 1 x x x4 + 1 x4 + 1

Thus

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 95

⎰ 3x dt 2
.0 ≤ lim (x + 2) · √ ≤ lim (x + 2) · √ = 0.
x→∞ x t t +1
4 x→∞ x +1
4

⎰ 3x dt
So, . lim (x + 2) · √ = 0.
x→∞ x t t4 + 1

S2007-4

(a) Reduce the integers modulo p, obtaining a uniform distribution over the ordered
list of Np remainders .(1, 2, . . . , p − 1, 0, 1, 2, . . . , p − 1, 0, 1, 2 . . . , p − 1, 0).
For any N, the probability a is divisible by p is the same as the probability of
selecting 0 from .(1, 2, . . . , p − 1, 0), which is . p1 .
(b) The selection of a and the selection of b are independent events. So, using part
(a), we see the probability is . p1 · p1 = p12 .
(c) Part (b) applies: .K = 2 · 3 · · · pk is of the form Np for any of the first k prime
numbers p. For each such prime p, the probability a and b both have p as a
factor is . p12 ; hence, the probability a and b are not both divisible by p is .1 − p12 .
Now, if .p1 and .p2 are distinct primes, whether or not .p1 divides both a and b
is independent from whether or not .p2 divides ⎛ a and ⎞b.⎛Hence, the ⎞ probability
that neither .p1 nor .p2 divides both a and b is . 1 − 1
1− 1
. Continuing,
p12 p22
we see that the probability that a and b have no common
⎛ factor
⎞ from among the
⎛ ⎞⎛ ⎞
1 1 1
first k primes is .P (k) = 1 − 2 1 − 2 · · · 1 − 2 . For sufficiently
2 3 pk
large k, this is arbitrarily close to both the .k → ∞ limit and the product over
∏⎛ 1
⎞
all primes, . 1− 2 .
p
p∈P

6
Remark from the Editors This infinite product is known to converge to .
π2
(see
[HWr]).

S2007-5 First, note that .A = B − I , where I is the .n × n identity matrix, and that
.B 2 = nB. For any real number r, we see .(B − I )(rB − I ) = rB 2 − (r + 1)B + I =
(rn − (r + 1))B + I , so .rB − I will be the inverse of .B − I if .rn − (r + 1) = 0, or
.r =
1
n−1 . Hence,
⎡ ⎤
· · · n−1
2−n 1
1
n−1 n−1
⎢ 1 ⎥
1 ⎢ · · · n−1
1 2−n
⎥
A−1 B −I =⎢ .. . . .. ⎥
n−1 n−1
. = ⎢ .. ⎥.
n−1 ⎣ . . . . ⎦
n−1 n−1 · · · n−1
1 1 2−n
30 2007 Solutions 97

S2007-6 Note that .h9 = (ghg −1 )3 = gh3 g −1 = g(ghg −1 )g −1 . So .h27 =

(ghg −1 )9 = gh9 g −1 = g(g 2 hg −2 )g −1 = g 3 hg −3 . Since .g 3 = e, .h27 = h, or
.h
26 = e. Since the group has odd order, the only possibilities for the order of h are

1 and 13. Since g and h do not commute, .h /= e; hence .|h| = 13.

Chapter 31
2008 Solutions

The 43rd competition was held at Saint Mary’s College.

The 2008 problem set was prepared by Professor Mike Axtell at Wabash College
and Professor Joe Stickles at Millikin University.
The 2008 problem statements begin on page 17 in Chap. 8.
S2008-1 We have

yx = (yx)3 = (yx)2 (yx) = ((yx)2 y)x

= y(yx)2 x = y(yx)(yx)x
= xy 2 yx 2 = xyx 2
= xx 2 y = x 3 y = xy.

Remark from the problem authors This problem and solution were both adapted
from ICMC P1974-2 [AFMC].

S2008-2 Consider the set .[0, 4] × [0, 4]. Draw in the line .y = x, which represents
both students meeting each other at the exact same time. If x arrives at time t, then
y can arrive at any time between .t − 0.5 and .t + 0.5, and the students will meet.
Similarly, if y arrives at time .t ' , then x can arrive at any time between .t ' − 0.5
and .t ' + 0.5, and the students will meet. So, widening the line .y = x so that it is
everywhere 1 hour wide and 1 hour tall forms a region that represents all possible
“successful” meeting times. The area of the .4 × 4-square represents all possible
selections of times by both students.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 99

The area of the shaded region is .16 − 2( 12 )( 72 )2 = 15

4 . So, the probability the two
15/4
students meet during this 4-hour period is . 16 = 64 .15

S2008-3 Let the angles of the triangles be denoted by .α, .β, .γ and .α ' , .β ' , .γ ' ,
and let .α = α ' . We ask under what conditions the following inequality holds:
' ' ' '
.sin α +sin β +sin γ < sin α +sin β +sin γ , or since .sin α = sin α , .sin β +sin γ <
' '
sin β + sin γ . By the hint, this inequality can be written as .2 sin( 2 ) cos( β−γ
β+γ
2 )<
' ' ' '
2 sin( β +γ β −γ ' ' ' ◦
2 ) cos( 2 ). Since .α = α , we have .β + γ = β + γ < 180 , so
β +γ ' '
that .2 sin( β+γ β+γ
2 ) = 2 sin( 2 ) > 0. Dividing both sides by .2 sin( 2 ) yields
β −γ ' '
.cos( β−γ
2 ) < cos( 2 ). Since .cos θ = cos |θ | and .y = cos| 'θ is' | decreasing
| | on
◦ ◦ | β −γ | | β−γ |
.0 ≤ θ ≤ 180 , the above inequality holds if and only if .| 2 | < | 2 |, or
equivalently, .|β ' − γ ' | < |β − γ |.

Remark from the problem authors This problem and solution were both adapted
from Problem .#2 of the 1898 Eötvös Competition in Hungary [K1 ].
√
S2008-4 The cross-sectional area is . 12 π( 2yi )2 = πyi at level .yi . The volume
of one puffed-up cross-section
∑ is .πyi Δyi , and we sum these together to get an
approximate volume of . πyi Δyi . This quickly leads to the exact volume given
⎰2
by . 0 πy dy = 2π .
S2008-5 First, we notice that

xn−1 + (n − 1)xn−2
xn =
.
n
nxn−1 − (n − 1)xn−1 + (n − 1)xn−2
=
n
−(n − 1)xn−1 + (n − 1)xn−2
= xn−1 + ,
n
31 2008 Solutions 101

or .xn − xn−1 = − n−1 n (xn−1 − xn−2 ). Similar arguments show .xn−1 − xn−2 =
− n−1 (xn−2 −xn−3 ), .xn−2 −xn−3 = − n−3
n−2
n−2 (xn−3 −xn−4 ), . . . , .x2 −x1 = − 2 (x1 −x0 ).
1

Substituting, we see for .n ≥ 2,

⎛ ⎞ ⎛ ⎞ ⎛ ⎞ ⎛ ⎞
n−1 n−2 n−3 1
xn − xn−1 = −
. · − · − · ... · − (x1 − x0 )
n n−1 n−2 2
(−1)n−1
= .
n
We also notice that the above equation is true when .n = 1. Since .x0 = 0, we can
write
xn = xn − x0
.

= (xn − xn−1 ) + (xn−1 − xn−2 ) + · · · + (x1 − x0 )

⎲
n
= (xk − xk−1 )
k=1

⎲
n
(−1)k−1
=
k
k=1

⎲
n−1
(−1)k
= .
k+1
k=0

∞
⎲ (−1)k
Hence . lim xn = = ln 2.
n→∞ k+1
k=0
S2008-6 We have
⎛ ⎞
1 2n 1 (2n)! (2n)!
. = · =
n+1 n n + 1 (n!)2 (n + 1)!n!
(2n)!(2n + 1 − 2n) (2n)!(2n + 1) (2n)!(2n)
= = −
(n + 1)!n! (n + 1)!n! (n + 1)!n!
(2n + 1)! 2n (2n)!
= − ·
(n + 1)!n! n (n + 1)!(n − 1)!
(2n + 1)! (2n)!
= −2·
(n + 1)!((2n + 1) − (n + 1))! (n + 1)!(2n − (n + 1))!
⎛ ⎞ ⎛ ⎞
2n + 1 2n
= −2 .
n+1 n+1
102 31 2008 Solutions

( ) ( 2n ) ( )
As both . 2n+1 1 2n
n+1 and . n+1 are integers for all integers .n ≥ 1, we have . n+1 n is an
integer for all integers .n ≥ 1 as well.
⎾ ⏋
1 −1
S2008-7A Let .A = . Then the characteristic polynomial of A is .λ2 −6λ+5.
0 5
By the Cayley–Hamilton Theorem, we have .A2 − 6A + 5I = 0. Multiplying both
sides by A, we obtain .A3 − 6A2 + 5A = 0. Now consider .(A − 2I )3 . We have
.(A − 2I ) = A − 6A + 12A − 8I = A − 6A + 5A + 7A − 8I = 7A − 8I .
3 3 2 3 2
⎛ ⎞3 ⎛ ⎞3
So, we have .7A = (A − 2I )3 + 8I , or .A = √ 3 (A − 2I )
1
+ √ 2
3 I . Thus, let
⎾ 1 ⏋ 7 ⎾ 2 ⏋ 7
−√ 3 −√ 3
1 √
3 0
.B = √ 3 (A − 2I ) = and .C = √3 I =
1 7 7 2 7
3 2 .
7 0 √
3
7 0 √ 3
7 7

S2008-7B
⎾ The
⏋ sum .B 3 + C 3
⎾ ⏋ is upper triangular, so try the upper triangular matrices
ab 00
.B = and .C = . Expand to get
0c 00
⎾ ⏋ ⎾ ⏋
a 3 a 2 b + abc + bc2 1 −1
B3 + C3 = B3 =
. = ,
0 c3 0 5

and let .a = 1, .c = 51/3 . Equating the upper right entries

b + 51/3 b + 52/3 b = −1,

there is a solution for b, which gives an example of the required matrices:

⎾ ( )⏋ ⎾ ⏋
1 −1/ 1 + 51/3 + 52/3 00
.B = , C= .
0 51/3 00
Chapter 32
2009 Solutions

The 44th competition was held at Indiana University—Purdue University Indi-

anapolis.
The 2009 problem set was prepared by Professor Chris Mitchell at the host
institution.
The 2009 problem statements begin on page 19 in Chap. 9.

S2009-1 The area between the lines .P = Q + 13 and .P = Q − 13 that is contained

in the square .[0, 2] × [0, 2] is equal to .4 − ( 35 )2 = 11
9 . Divide this number by 4 to
get the required probability.

S2009-2 Consider the equation

⎲ ∞
x
. = fn x n .
1 − x − x2
n=1

Multiply both sides of this equation by the .1 − x − x 2 , and expand the right-hand
side of the equation. The result, after reindexing, is the equation
∞
⎲
x = f1 x + (f2 − f1 )x 2 +
. (fn − fn−1 − fn−2 )x n .
n=3

The result follows by comparing coefficients on both sides of the equation.

S2009-3 Note that, using cycle notation, .σ 3 = (1 2 3 5 7 8 4 6). Because .σ 3 can

be written as a single 8-cycle, .σ must also be writable as some 8-cycle (by the
uniqueness of expressing permutations as a product of disjoint cycles). Because all
8-cycles have an order of 8, it follows that .σ = σ 9 = (σ 3 )3 = (1 5 4 2 7 6 3 8).

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 103
A. Coffman et al., The Indiana College Mathematics Competition (2001–2023),
Problem Books in Mathematics, https://doi.org/10.1007/978-3-031-62768-2_32
104 32 2009 Solutions

√3 √ √3 √
S2009-4 Let .v = 2 + 5 + 2 − 5. Note that .v ∈ R, and expanding .v 3 shows
that v satisfies the equation .x 3 +3x −4 = 0. Since .x 3 +3x −4 = (x −1)(x 2 +x +4)
and the discriminant of .x 2 + x + 4 is .−15, the only real solution of .x 3 + 3x − 4 = 0
is 1. So, .v = 1.

Remark from the Editors This problem is similar to P2014-8—the solution S2014-8
in Chap. 37 shows a more detailed calculation.

S2009-5 A stronger result holds under the given conditions: .n|(n−3)!. A composite
number .0 < n /= 4 satisfies .n ≥ 6 and .n = a · b for .1 < a ≤ b ≤ 12 n ≤ n − 3. If
.a < b, then a and b are distinct factors of .(n − 3)!, so .n = ab divides .(n − 3)!. If

.a = b, then .2 < a < 2a ≤ (a − 1)a = a − a ≤ n − 3, so a and 2a are distinct

factors of .(n − 3)! and .n = a divides .(n − 3)!.

Remark from the Editors This well-known problem, showing that Wilson’s Theo-
rem only works for primes, is considered, for example, in [RL].

S2009-6A Step 1. For any three points in the interval, .a < r < s < t < b, the
inequality

f (λx + (1 − λ)y) ≤ λf (x) + (1 − λ)f (y)

t −s
applied to .x = r, .y = t, .λ = ∈ (0, 1) gives
t −r
⎛ ⎛ ⎞ ⎞ ⎛ ⎞
t −s t −s t −s t −s
.f r + 1− t ≤ f (r) + 1 − f (t)
t −r t −r t −r t −r
t −s s−r
=⇒ f (s) ≤ f (r) + f (t).
t −r t −r

This implies (without assuming f is increasing)

⎛⎛ ⎞ ⎞
f (s) − f (r) t −s s−r
. ≤ f (r) + f (t) − f (r) /(s − r)
s−r t −r t −r
f (t) − f (r)
=
t −r
⎛ ⎛ ⎞⎞
t −s s−r
= f (t) − f (r) + f (t) /(t − s)
t −r t −r
f (t) − f (s)
≤ .
t −s
32 2009 Solutions 105

Step 2. To show that f is continuous at a particular point .x0 ∈ (a, b), let .ϵ > 0, and
f (t0 ) − f (x0 )
pick any point .t0 ∈ (x0 , b). Let .m = > 0 (using f increasing), and
t0 − x0
let .δ = min{ϵ/(2m), t0 − x0 , x0 − a} > 0. If .0 < x − x0 < δ, then .a < x0 < x <
t0 < b and, using Step 1,

f (x) − f (x0 )
0 < f (x) − f (x0 ) =
. · (x − x0 )
x − x0
f (t0 ) − f (x0 )
≤ · δ ≤ m · (ϵ/(2m)) < ϵ.
t0 − x0

If .0 < x0 − x < δ, then .a < x < x0 < t0 < b, and using Step 1,

f (x0 ) − f (x)
0 < f (x0 ) − f (x) =
. · (x0 − x)
x0 − x
f (t0 ) − f (x0 )
≤ · δ ≤ m · (ϵ/(2m)) < ϵ.
t0 − x0

The conclusion is that .|f (x) − f (x0 )| < ϵ, so . lim f (x) = f (x0 ) and f is
x→x0
continuous.

S2009-6B A stronger conclusion is possible, even without the increasing assump-

tion: Both one-side derivatives of f exist at each point (but are not necessarily equal,
e.g., .f (x) = |x|).
To show that f has a right-side derivative at a particular point .x0 ∈ (a, b), consider
the set
⌠ ⎫
f (x) − f (x0 )
. : x0 < x < b .
x − x0

This set is non-empty and bounded below: For some .r0 ∈ (a, x0 ),

f (x) − f (x0 ) f (x0 ) − f (r0 )

. ≥
x − x0 x0 − r0

by Step 1. So, the set has a greatest lower bound, L, and for any .ϵ > 0, there is some
t0 ∈ (x0 , b) such that
.

f (t0 ) − f (x0 )
L≤
. < L + ϵ.
t0 − x0

If .0 < x − x0 < t0 − x0 , then by Step 1,

f (x) − f (x0 ) f (t0 ) − f (x0 )

L≤
. ≤ < L + ϵ.
x − x0 t0 − x0
106 32 2009 Solutions

This shows
f (x) − f (x0 )
. lim = L = f+' (x0 ).
x→x0+ x − x0

The right-side continuity follows from the usual calculation:

⎛ ⎞
f (x) − f (x0 )
. lim (f (x) − f (x0 )) = lim · (x − x0 ) = f+' (x0 ) · 0 = 0.
x→x0+ x→x0+ x − x0

In the same way, the left-side derivative .f−' (x0 ) is the least upper bound of the set
of left-side difference quotients. The left-side continuity and the two-side continuity
follow.

S2009-7A There may be many such bijections from .(Z≥0 )3 to .Z≥1 —the following
example is constructed using the increasing sequence of triangular numbers

.(0, 1, 3, 6, 10, . . . , Ta , . . .),

defined by .Ta = 12 a(a + 1), or recursively by .T0 = 0, and .Ta+1 = Ta + a + 1, and

another increasing sequence, the tetrahedral numbers,

. (0, 1, 4, 10, 20, 35, 56, 84, 120, 165, 220, 286, 364, 455,
560, 680, 816, 969, 1140, 1330, 1540, 1771, 2024, . . . , Eb , . . .),

defined by .Eb = 16 b(b+1)(b+2), or recursively by .E0 = 0 and .Eb+1 = Eb +Tb+1 .

Then define .F : (Z≥0 )3 → Z≥1 by

F (p, q, r) = 1 + p + Tp+q + Ep+q+r .

To show F is surjective, consider .x ≥ 1, and let .b = max{y ∈ Z : Ey < x}, so that

Eb < x ≤ Eb+1 and .0 < x − Eb ≤ Eb+1 − Eb = Tb+1 . Let .a = max{y ∈ Z : Ty <
.

x − Eb }, so that .Ta < x − Eb ≤ Ta+1 and .0 < x − Eb − Ta ≤ Ta+1 − Ta = a + 1.

From .Ta < x − Eb ≤ Tb+1 , we can conclude .a < b + 1, so .b − a ≥ 0. Then one
can verify that

.F (x − Eb − Ta − 1, a − (x − Eb − Ta − 1), b − a) = x.

To show F is injective, suppose there are two input triples; label the pair with the
smaller or equal sum of components by .(p, q, r) and the other pair by .(p∗ , q ∗ , r ∗ ),
so .p + q + r ≤ p∗ + q ∗ + r ∗ .
We want to show that assuming the outputs are equal
32 2009 Solutions 107

F (p, q, r) = F (p∗ , q ∗ , r ∗ )
.

1 + p + Tp+q + Ep+q+r = 1 + p∗ + Tp∗ +q ∗ + Ep∗ +q ∗ +r ∗ (32.1)

leads to the conclusion .p = p∗ and .q = q ∗ and .r = r ∗ .

Case 1. p + q + r < p∗ + q ∗ + r ∗ . Then, using the increasing property of the .Eb

and .Ta sequences, (32.1) implies

0 = Ep∗+q∗+r ∗ − Ep+q+r + Tp∗ +q ∗ − Tp+q + p∗ − p

≥ Ep+q+r+1 − Ep+q+r − Tp+q + p∗ − p

= Tp+q+r+1 − Tp+q + p∗ − p
≥ Tp+q+1 − Tp+q + p∗ − p
= p + q + 1 + p∗ − p
= p∗ + q + 1 ≥ 1,

a contradiction.
Case 2. p + q + r = p∗ + q ∗ + r ∗ and .p + q < p∗ + q ∗ . From (32.1),
.

0 = Tp∗ +q ∗ − Tp+q + p∗ − p
.

≥ Tp+q+1 − Tp+q + p∗ − p
= p + q + 1 + p∗ − p
= p∗ + q + 1 ≥ 1,

a contradiction.
Case 3. p + q + r = p∗ + q ∗ + r ∗ and .p + q > p∗ + q ∗ . Then (32.1) leads to a
.

contradiction in the same way as Case 2.

Case 4. The only remaining case is .p+q +r = p∗ +q ∗ +r ∗ and .p+q = p∗ +q ∗ ,
so .r = r ∗ follows immediately, and then .p = p∗ follows from (32.1), and
so in this case, .q = q ∗ .

Remark from the Editors There may be descriptions, sketches, or algorithms

describing such a correspondence, generalizing the better-known diagonal con-
struction of a map .(Z≥0 )2 to .Z≥1 from two dimensions to three, but the above
construction gives a concrete formula as stated in the problem.

S2009-7B If an explicit formula is known for a bijection .g : (Z≥0 )2 → Z≥1 (as in

P2022-8, see the solution S2022-8 in Chap. 44), then .F (p, q, r) = g(g(p, q)−1, r)
is a bijection because it is a composite of bijections.
Chapter 33
2010 Solutions

The 45th competition was held at Franklin College.

The 2010 problem set was prepared by Professor Chris Mitchell at American
University.
The 2010 problem statements begin on page 21 in Chap. 10.
( )
S2010-1 The solution is . p+q
q . In order to see this, think of .p + q boxes. Each box
will contain a card. The deck is divided into two parts, one containing p cards and
one containing q cards. Shuffling the two parts together is the same as selecting q
of the .p + q boxes and then placing the cards in order into those boxes.

S2010-2 First, recall Wilson’s Theorem (see [HWr], [M], or [RL]):

(p − 1)! ≡ −1 (mod p).

Note that for every .0 < i ≤ k, .p − i ≡ −i (mod p). As a result

.(p − (k + 1))!k! ≡ (−1)k (p − (k + 1))!(p − k) · · · (p − 1) (mod p).

S2010-3 Player 1 should take 1, 2, or 3 coins at any stage of the game, depending
on the number of coins left being equivalent to 2, 3, or 0 modulo 4.
Let n be the number of coins at a point in the game when it is player 1’s turn to act,
but player 1 has not yet taken the coins (such as at the start of the game). We can
write n as .n = 4k + i for .i = 0, 1, 2, 3.
First consider the following cases where .k = 0, so .n = i:

• If .i = 2, player 1 will win by choosing 1 coin.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 109
A. Coffman et al., The Indiana College Mathematics Competition (2001–2023),
Problem Books in Mathematics, https://doi.org/10.1007/978-3-031-62768-2_33
110 33 2010 Solutions

• If .i = 3, player 1 will win by choosing 2 coins.

• If .i = 0, player 1 has won because player 2 has already taken the last coin.

The next step is induction on k. Assume that it is known that if there are .4p +i coins
left on the table when it is player 1’s turn to act, with .0 ≤ p ≤ k, and .i = 0, 2, or 3,
then player 1 wins. Suppose that .i = 0, 2, or 3 and that there are .n = 4(k + 1) + i
coins on the table, so there are at least 4 coins, and it is player 1’s turn to act. Player
1’s strategy is to take .i − 1 modulo 4 coins (so if .i = 0, then player 1 takes 3 coins).
Then, player 2 will take j coins, where j is equal to 1, 2, or 3. The remaining number
of coins is then equivalent modulo 4 to

. (4(k + 1) + i) − (i − 1) − j ≡ 1 − j (mod 4),

which is .1 − 1 ≡ 0, .1 − 2 ≡ 3, or .1 − 3 ≡ 2 modulo 4. In particular,

• If .i = 0 and .j = 2 or 3, then there are .4(k + 1) − 3 − j = 4(k − 1) + (5 − j )

coins left.
• If .i = 0 and .j = 1, then there are .4(k + 1) − 3 − 1 = 4k + 0 coins left.
• If .i /= 0 and .j = 2 or 3, then there are .4(k + 1) + i − (i − 1) − j = 4k + (5 − j )
coins left.
• If .i /= 0 and .j = 1, then there are .4(k + 1) + i − (i − 1) − 1 = 4(k + 1) coins
left.

In the first three cases, the number of coins left is .4p + i ' , with .p = k − 1 or k, and
'
.i = 0, 2, or 3, so by the inductive hypothesis, player 1 will win. In the last case, the

number of coins left is .4(k + 1) + 0 ≥ 4, so the game is not over and the next moves
by players 1 and 2 will fall into one of the first two cases where .i = 0, and then the
inductive hypothesis will apply.

S2010-4 If the rank of B is 0, then .B = 0, so .A = I follows from .A + B = I

and the required conditions hold. Otherwise, if the rank of B is p with .0 < p ≤ n,
then by the Rank + Nullity Theorem for A, the dimension of the nullspace of A is
p, and there is an ordered basis for the nullspace of A, .(→ a1 , . . . , a→p ). For each of
ak ) = (I − A)(→
the nullspace vectors, .B(→ ak ) = a→k − 0→ = a→k . So, .(→ a1 , . . . , a→p ) is an
independent list of vectors in the p-dimensional image of B and must be a basis of
the image of B. For any .x→, .A(B(→
x )) = A(c1 a→1 + . . . + cp a→p ) = 0.→ From .AB = 0,
we can conclude .A(I − A) = 0 =⇒ A − A2 = 0 =⇒ A = A2 , and .(I − B)B =
0 =⇒ B −B 2 = 0 =⇒ B = B 2 . Finally, .BA = (I −A)A = A−A2 = 0 = AB.

S2010-5 The quotient is not defined at .x = 0. The proof will proceed by

noting the Taylor series expansions of the various functions and composing these
appropriately. The Taylor series for the relevant functions are
33 2010 Solutions 111

x3
. sin x = x − + Terms of at least order 5,
3!
x3
tan x = x + + Terms of at least order 5,
3
x3
arcsin x = x + + Terms of at least order 5,
3!
x3
arctan x = x − + Terms of at least order 5.
3
After composition and some manipulation, we obtain the following limit:

−x 3 + Terms of at least order 5

. lim
x→0 x 3 + Terms of at least order 5

As a consequence, the limit is .−1.

S2010-6 The thing to observe is that we are looking for ordered pairs .(x, y) in
the first quadrant that satisfy the inequality . 12 < yx < 3 whose sum is odd and
bounded above by 5. There are only three ordered pairs that satisfy these conditions:
.(1, 2), .(2, 3), and .(3, 2). .U + D + R + L = 5, where the letters represent the

number of moves up (U ), down (D), and so forth. Let the ordered quadruple
.(U, D, R, L) represent the set of directions. To get to .(2, 3), .U = 3 and .R = 2.
5! 1 2 1 3
The associated probability is . 2!3! ( 16 ) ( 4 ) . Similarly, the probability associated
5! 1 3 1 2
with the point .(3, 2) is . 3!2! ( 16 ) ( 4 ) . Both .(2, 0, 2, 1) and .(3, 1, 1, 0) will reach
5! 1 2 1 2 9
.(1, 2). The probability associated with the first quadruple is .
2!2!1! ( 4 ) ( 16 ) ( 16 ).
5! 1 3 1 1
The probability associated with the second quadruple is . 3!1!1! ( 4 ) ( 8 )( 16 ). The sum
of these four probabilities is the solution.

S2010-7 The difficulty is in finding a way to show that for rational numbers close
to a given irrational number x, the denominator gets large. We may also assume that
.x ∈ [0, 1] and that x is irrational. Let .ϵ > 0. There is an .n ∈ N such that .
1
n < ϵ. For
every .1 ≤ i ≤ n, define
⌠ | | | | ⎫
| 1 || | i − 1 ||
.δi = min |x − 0|, |x − |
, . . . , |x − , |x − 1| .
| i| i |

Define .δ = 12 min{δi }; it follows from the irrational property of x that .δ > 0. The
claim is that if .|y −x| < δ, then .|f (y)−f (x)| < ϵ, so by definition, f is continuous
at x. To prove the claim, consider .y ∈ (x − δ, x + δ).

Case 1. If .y ∈
/ Q, then .|f (x) − f (y)| = 0.
112 33 2010 Solutions

Case 2. If .y ∈ Q, with .y = pq in the lowest terms and .q ≤ n, then .|x − pq | < δ <
δq ≤ |x − pq |, a contradiction.
p
Case 3. The only remaining case is that .y ∈ Q, with .y = q in the lowest terms
and .q > n, so .|f (y) − f (x)| = | q1 − 0| < 1
n < ϵ.
Chapter 34
2011 Solutions

Indiana Wesleyan University hosted an Indiana MAA meeting and ICMC event for
the first time in 2011.
The 2011 problem set was prepared by Professor Chris Mitchell at American
University.
The 2011 problem statements begin on page 23 in Chap. 11.

S2011-1 No, there is no even number .n ≥ 6 such that it is possible to achieve any
elimination order by choosing a corresponding elimination parameter. Consider, for
any even number .n ≥ 6, the following order of elimination: .1, n, n−1, n−2, . . . , 3,
so that child 2 is the winner. If the first child on the list is person 1, then the parameter
i must be of the form nk, for some integer k; therefore the elimination parameter
must be even. If the game continues until only players 2, 3, and 4 remain, then
according to the specified order, child 4 leaves, and the count starts again with child
2. However, any even count starting with child 2 would end with child 3 and require
child 2 to leave, contradicting the chosen elimination order.

S2011-2 If x represents the length of the cut, let .P (x), .A(x), .S(x), and .V (x)
represent the perimeter of the base, the area of the base, the area of the sides, and
the volume of the resulting box, as functions of x.
Drawing a line segment from a vertex of the interior triangle to the corresponding
vertex of the exterior triangle creates two .30◦ − 60◦ − 90◦ triangles. Since the short
leg has√length x, we can conclude the hypotenuse has length 2x and the long leg √has
length . 3x. It follows that each side of the interior triangle has length .1 − 2 3x,
√ √ √
.P (x) = 3(1 − 2 3x), and .S(x) = 3x(1 − 2 3x). The lengths x and .1 − 2 3x are

nonnegative on the domain interval .[0, √ 1

].
2 3
Drawing an altitude from one vertex of the interior triangle to its opposite side yields
another pair of .30◦ − 60◦ − 90◦ triangles. Using a similar method as before, we can

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 113
A. Coffman et al., The Indiana College Mathematics Competition (2001–2023),
Problem Books in Mathematics, https://doi.org/10.1007/978-3-031-62768-2_34
114 34 2011 Solutions

√ √
conclude the altitude length (i.e., height) of the interior triangle is . 23 (1 − 2 3x)
√ √ √
and, thus, .A(x) = 43 (1 − 2 3x)2 . It immediately follows that .V (x) = 43 x(1 −
√ 2
2 3x) .
The result is implied by maximizing .V (x) on .[0, √ 1
] using the tools of calculus.
2 3
'
√ √
Note that .V (x) = 9 3x −6x+ 4 . Setting the derivative equal to 0 and solving for
2 3

x (using the quadratic formula) yields .x = √

1
or .x = √ 1
. The former obviously
2 3 6 3
produces a minimal volume because .V ( 1
√ ) = 0; the latter value for x uniquely
2 3
produces a maximal volume. The two lengths .x = 1
√ , .x = 1
√ , are also the only
2 3 6 3
solutions of .A(x) = S(x).
Finally, we note that .A( 1
√ ) = 1
√ = S( 1
√ ), which completes the problem.
6 3 3 3 6 3

S2011-3 For a particular number n, the function defined for .(x, y) /= (0, 0) by

x n+1 y
f (x, y) =
.
x 2(n+1) + y 2

(and then extended to the domain .R2 by .f (0, 0) = 0) satisfies

. lim f (x, g(x)) = 0

x→0

for all .g ∈ C n . Let .y = x n+1 ; then

x 2n+2 1
. lim f (x, y) = lim = /= 0.
x→0 x→0 x 2n+2 +x 2n+2 2

Thus, in particular, it is not true that

. lim f (x, y) = 0,
(x,y)→(0,0)

and in fact the limit does not exist.

Remark from the Editors The .n = 1 case of this example appears in ICMC P1986-6
[AFMC].

S2011-4 If n and b are relatively prime, then there is a minimal nonzero r for which
br ≡ 1 (mod n), or .br − 1 = kn, for some k. It follows that
.

k = dr−1 br−1 + · · · d1 b + d0 for 0 ≤ di < b.

Putting the last two ideas together, we obtain

34 2011 Solutions 115

1 dr−1 br−1 + · · · d1 b + d0 1
. = .
n b r 1 − b−r

⎲ ∞
1
Since . = (b−r )i , a cycle length for . n1 is r. To show that there is no
1 − b−r
i=0
lower cycle length, note that the equation (which would follow if m were another
cycle length)

1 em−1 bm−1 + · · · e1 b + e0 1
. =
n bm 1 − b−m

implies that .m ≥ r because of the minimality of r.

S2011-5 Imagine that the three pegs are occupied, from left to right, by the red
stack and then the blue stack, and the rightmost peg is unoccupied. Label these three
pegs .PL , .PM , and .PR (for left peg, middle peg, and right peg). In order to swap the
positions of the two stacks, we will create a double stack containing .2(n − 1) disks
of alternating color (the pattern will be, from the top; red, blue, red, blue, and so on)
over the largest blue disk on .PM . We will then be able to move: 1. the largest red
disk to .PR ; 2. the double stack over the largest red disk on .PR ; 3. the largest blue
disk onto .PL ; 4. the double stack onto the largest blue disk on .PL ; 5. the largest red
disk onto .PM . Finally, we unstack the double stack. Let .DM (k) denote the number
of moves required to create a double stack of 2k disks on .PM . Let .A(k) denote the
number of moves required to implement the entire algorithm. Since it will require
.2(2 − 1) moves to shift the double stack from one peg to another, we would have
k

the following:

. A(n) = 2DM (n − 1) + 3 + 4(2n−1 − 1).

Suppose that one has two stacks of k red and blue disks. Let .DR (k) denote the
number of moves required to shuffle these two stacks into a single double stack
which sits atop .PR . In order to create a double stack of 2k disks over the .PM , one
is required to create a double stack of .2(k − 1) disks over .PR . Once this has been
done, the second largest red disk would be moved atop the second largest blue disk
on .PM , and then the double stack of .2(k − 1) disks would be moved from .PR onto
.PM . We would then have

DM (k) = DR (k − 1) + 2(2k−2 − 1) + 1.
.

On the other hand, in order to create a double stack of 2k disks on .PR , we are
required to create a double stack of .2(k − 1) disks on .PM , move a red disk from .PL
peg onto .PR , move the double stack of .2(k − 1) disks onto the leftmost peg, move
a blue disk from the center peg onto the rightmost peg, and then move the double
stack onto the rightmost peg. So:
116 34 2011 Solutions

DR (k) = DM (k − 1) + 2 + 4(2k−1 − 1).

Putting these two together, we would have the relation:

DM (k) − DM (k − 2) = 3 + 2(2k−2 − 1) + 4(2k−2 − 1).

Since .DM (0) = 0 and .DM (1) = 1, this relation permits an efficient calculation of
DM (k) for any k by examining the telescoping sum (for .i = 2 or .i = 3):
.

. (DM (k) − DM (k − 2)) + (DM (k − 2) − DM (k − 4)) + · · ·

· · · + (DM (i) − DM (i − 2))
= DM (k) − DM (i − 2).

For even k write .k = 2j . Then

⎲
j
DM (2j ) − DM (0) =
. (DM (2i) − DM (2i − 2))
i=1
j ⎛
⎲ ⎞
= 3 · 22i−1 − 3
i=1

4j +1 − 4
=3· − 3j
2·3
= 2 · 22j − 2 − 3j
3
= 2 · 2k − 2 − k.
2

For odd k write .k = 2j + 1. Then

⎲
j
DM (2j + 1) − DM (1) =
. (DM (2i + 1) − DM (2i − 1))
i=1
j ⎛
⎲ ⎞
= 3 · 22i − 3
i=1

4j +1 − 4
=3· − 3j
3
= 22j +2 − 4 − 3j
34 2011 Solutions 117

5 3
= 2 · 2k − − k.
2 2

Thus .DM (2j + 1) = DM (1) + 2 · 2k − 52 − 23 k = 2 · 2k − 3

2 − 32 k.
⌠
2k+1 − 2 − 32 k if k is even,
Therefore .DM (k) =
2k+1 − 32 − 23 k if k is odd.
7+(−1)k
Alternatively this may be written as .DM (k) = 2k+1 − 32 k − 4 . Substitute this
into the formula for .A(n) to get

A(n) = 2DM (n − 1) + 3 + 4(2n−1 − 1)

3 − (−1)n
= 2n+2 − 3n − .
2

S2011-6 Assume that .xn /= f (xn ) for all .n ≥ 1. Otherwise, a fixed point clearly
exists, and the limit converges to this fixed point. In particular, .x1 /= f (x1 ). By the
Mean Value Theorem, there is some .x1∗ between .x1 and .f (x1 ) = x2 so that

f (x2 ) − f (x1 )
. = f ' (x1∗ )
x2 − x1
|f (x2 ) − f (x1 )| 1
=⇒ = |f ' (x1∗ )| ≤
|x2 − x1 | 2
|x2 − x1 |
=⇒ |x3 − x2 | ≤ .
2
More generally for .n > 1, using .xn /= xn−1 ,

|x2 − x1 |
|xn+1 − xn | ≤
. (34.1)
2n−1
for all .n ≥ 1. For .n > m + 1 ≥ 2, using the triangle inequality and (34.1):

|xn − xm | ≤ |xn − xn−1 | + |xn−1 − xn−2 | + · · · + |xm+1 − xm |

.
⎛ ⎞
1 1 1
≤ |x2 − x1 | · + + · · · +
2n−2 2n−3 2m−1
118 34 2011 Solutions

|x2 − x1 |
< .
2m−2
|x2 − x1 |
To summarize, the inequality .|xn − xm | < holds for all .n ≥ m ≥ 1. So,
2m−2
|x2 − x1 |
for any .ϵ > 0, choose .N ≥ 1 so that . N −2 < ϵ. Then for .n ≥ m ≥ N,
2
|x2 − x1 | |x2 − x1 |
|xn − xm | <
. ≤ < ϵ.
2m−2 2N −2
This implies that the sequence is Cauchy, therefore convergent. Since the limit exists
and f is continuous, it must be that

f ( lim xn ) = lim f (xn ) = lim xn .

.
n→∞ n→∞ n→∞

S2011-7 The following function .f : R → R \ {0} is a one-to-one correspondence:

⌠
x if x is not a whole number,
.f (x) =
x + 1 if x is a whole number.

S2011-8 The matrix in question can be thought of as the evaluation of a polynomial

with complex coefficients at the following matrix X:
⎡ ⎤
00 0 1···
⎢1 0 0 0⎥
···
⎢ ⎥
⎢ 0⎥
···
.⎢0 1 0 ⎥.
⎢. . .. .. ⎥
..
⎣ .. .. . .⎦ .
0 0 ··· 1 0

To be more precise, if A is the matrix described in the problem and I is the identity
matrix, then

A = I + nX + (n − 1)X2 + · · · + 2Xn−1 .
.

It follows that if v is an eigenvector for X, then v is an eigenvector for A. From

.det(X − λI ) = ±(1 − λn ), we see that X has n distinct complex eigenvalues (the
.nth roots of unity), and therefore there exists a list of n eigenvectors for X which is

linearly independent (over .C). Such a list of eigenvectors for X is a complete list of
eigenvectors for A. Any eigenvector v for X is a (nonzero) complex scalar multiple
of a vector of the form
34 2011 Solutions 119

⎡ ⎤
1
⎢ ωk ⎥
⎢ ⎥
⎢ ωk2 ⎥
.⎢ ⎥,
⎢ .. ⎥
⎣ . ⎦
ωkn−1

where .ωk is one of the .nth roots of unity, .{ω0 , ω1 , . . . , ωn−1 } (including .ω0 = 1).
Chapter 35
2012 Solutions

The .47th competition was held at Ball State University.

The 2012 problem set was prepared by Professor Ahmed Mohammed at the host
institution.
The 2012 problem statements begin on page 27 in Chap. 12.

S2012-1 Obviously, the statement is true for .n = 1. So we assume that .n ≥ 2. By

the Binomial Theorem we note that
⎛ ⎞
⎲n ⎛ ⎞ ⎲ n ⎛ ⎞
⎝ n j⎠ n j
.(n + 1) − 1 = n −1=
n
n
j j
j =0 j =1
⎛ ⎞ n ⎛ ⎞
⎲ n ⎛ ⎞
⎲
n n j n j
= n+ n =n +
2
n .
1 j j
j =2 j =2

( )
On noting that . nj is a positive integer, we see that

n ⎛ ⎞
⎲ n j
. n
j
j =1

is divisible by .n2 . Therefore, indeed .(n + 1)n − 1 is divisible by .n2 .

S2012-2 First we observe that the number of zeros at the end of .213! is same as the
number of times the number 10 occurs as a factor of .213!. More precisely, since
.10 = 2 × 5, and since there are more factors of 2 than 5 in the unique prime

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 121
A. Coffman et al., The Indiana College Mathematics Competition (2001–2023),
Problem Books in Mathematics, https://doi.org/10.1007/978-3-031-62768-2_35
122 35 2012 Solutions

factorization of the product .213!, we observe that there are as many ending zeros
in the product .213! as there are factors of 5. The number of positive multiples of 5
less or equal to 213 is
⎡ ⎤
213
. = 42, (35.1)
5

where .[x] denotes the greatest integer less than or equal to x. Among these
multiples, we list those that contain two or more factors of 5 as follows:

25 = 52 ,
. 50 = 2 · 52 , 75 = 3 · 52 , 100 = 4 · 52
125 = 53 , 150 = 6 · 52 , 175 = 7 · 52 , 200 = 8 · 52 .

Therefore there are exactly 9 additional occurrences of the digit 5 that have not been
counted in (35.1). The conclusion is that there are a total of .42+9 = 51 occurrences
of the digit 5 in the product .213!, and so 51 zeros at the end of .213!.

Remark from the Editors For a general approach to finding the number of times a
prime p occurs as a factor of a factorial .m!, see the solution of Problem .#2 of the
1925 Eötvös Competition in Hungary [K2 ].

S2012-3

(a) Suppose .r/s is a rational root so that r and s are integers with .s /= 0. Clearly
.r /= 0. We suppose that .gcd(r, s) = 1, so that the fraction .r/s is in lowest terms,

and in particular, r and s cannot both be even. We proceed to show that all other
parities lead to a contradiction.
By assumption we have
⎛r ⎞ ⎛ r ⎞n ⎛ r ⎞2 ⎛r ⎞
0=p
. = an + · · · + a2 + a1 + a0 .
s s s s
Clearing fractions, we see that

an r n + an−1 sr n−1 + · · · a2 s n−2 r 2 + a1 s n−1 r + a0 s n = 0.

. (35.2)

Case 1. Suppose either r or s is even and the other is odd. This would imply
that every LHS term of (35.2) except one is even, with the remaining
term odd, contradicting RHS being even.
Case 2. Suppose now both s and r are odd. Add the odd integer .an + · · · + a2
to both sides of (35.2) and get
35 2012 Solutions 123

. an (1 + r n ) + an−1 (1 + sr n−1 ) + · · · + a2 (1 + s n−2 r 2 ) + a1 s n−1 r

+a0 s n
= an + · · · + a2 . (35.3)

Then .1 + s i r j is an even integer for any nonnegative integers .i, j . Therefore

.an (1 + r ) + an−1 (1 + sr
n n−1 ) + · · · + a (1 + s n−2 r 2 ) is even. Since .a and .a
2 0 1
are both odd, we note that each of .a1 s n−1 r and .a0 s n is odd, and therefore their
sum is even. Therefore we see that LHS of (35.3) is even. But we recall that
RHS is odd. Thus, once again, we conclude that r and s cannot be both odd.
(b) Each of the following polynomials has a rational root:

.p1 (x) = 1x 2 + 1x − 2 =⇒ p1 (1) = 0

p2 (x) = 1x 2 − 2x + 1 =⇒ p2 (1) = 0
1
p3 (x) = 2x 3 + 1x 2 + 1x − 1 =⇒ p3 ( ) = 0
2
p4 (x) = 1x 3 + 1x 2 − 1x − 1 =⇒ p4 (1) = 0.

These examples show that none of the conditions in the problem may be
omitted.

Remark from the Editors For the polynomial .p(x) = an x n +. . .+a0 with a rational
root .r/s in lowest terms, the Rational Root Theorem (see S2001-4A in Chap. 24)
states that r is a divisor of .a0 and s is a divisor of .an . So if .an and .a0 are both odd,
then r and s are also both odd; the above Case 1. argument is a special case of this
idea.

S2012-4A First let us dispose of the trivial case when .β = 0. In this case .A = αIn ,
which is obviously diagonalizable, and .det A = α n . So, henceforth we suppose that
.β /= 0. Note that .A = w→
→ v T − γ In , where
⎡ ⎤ ⎡ ⎤
1 β
⎢1⎥ ⎢β ⎥
⎢ ⎥ ⎢ ⎥
→ = ⎢.⎥ ,
.w v→ = ⎢ . ⎥ , and γ = β − α.
⎣ .. ⎦ ⎣ .. ⎦
1 β

Now, a nonzero vector .x→ in .Rn is an eigenvector of A if and only if .A→

x = λ→
x for
some .λ ∈ R. That is,

→ v T x→ − γ x→ = λx.
w→
.

We rewrite this as
124 35 2012 Solutions

v T x→)w
(→
. → = (λ + γ )→
x.

If .c = v→T x→ /= 0, then note that .λ + γ /= 0, and therefore

x→ = c(λ + γ )−1 w,
. →

showing that .x→ is a multiple of .w.

→ If .v→T x→ = 0, then .x→ is orthogonal to .v→ and hence
→ (recall that .β /= 0). Thus, any eigenvector of A is either a multiple of .w
to .w → or
→ Therefore we see that .w
orthogonal to .w. → is an eigenvector of A with

→ = w→
Aw
. → vT w
→ − γw
→ = (nβ − γ )w
→ = ((n − 1)β + α)w.
→

That is, .λ = (n − 1)β + α is an eigenvalue of A with corresponding eigenspace of

dimension 1. On the other hand, any other eigenvector of A must be orthogonal
→ with corresponding eigenvalue .λ = −γ = α − β. The eigenspace of A
to .w
corresponding to .λ = α − β is the orthogonal complement of the eigenspace of
A corresponding to the eigenvalue .λ = (n − 1)β + α, namely the orthogonal
→ Thus the eigenspace of A corresponding to
complement of the line parallel to .w.
the eigenvalue .λ = α − β has dimension .n − 1. This shows that A is diagonalizable.
The determinant of A is the product of its eigenvalues

. det A = (α − β)n−1 ((n − 1)β + α) .

S2012-4B As in the previous solution, let .w → = [1, 1, . . . , 1]T , so that .A = (α −

β)In + β w→w→ . Let .e→1 = [1, 0, . . . , 0] be the first standard basis column vector,
T T

and let P be the invertible matrix where the first column of .In is replaced by .w: →
⎡ ⎤
1
⎢1 1 ⎥
⎢ ⎥
.P = ⎢ . . ⎥,
⎣ .. .. ⎦
1 1

→ and .P −1 w
so that .P e→1 = w → = e→1 . Then A is similar to an upper triangular matrix:

P −1 AP = (α − β)In + βP −1 w
. →w→T P
= (α − β)In + β e→1 [n, 1, . . . , 1]
⎡ ⎤
α + (n − 1)β β · · · β
⎢ α−β ⎥
⎢ ⎥
=⎢ .. ⎥.
⎣ . ⎦
α−β
35 2012 Solutions 125

By inspection, .det(P −1 AP ) = (α − β)n−1 (α + (n − 1)β), .e→1 is an eigenvector of

.P
−1 AP with eigenvalue .α + (n − 1)β, and there are .n − 1 eigenvectors of the form

.[1, 0, . . . , 0, −n, 0, . . . , 0] with eigenvalue .α − β. So, .P

T −1 AP has n independent

eigenvectors and is diagonalizable, and it follows that A is also diagonalizable with

the same determinant.

S2012-5 Let .N = {e, a} be a normal subgroup of G of order 2, where e is the

identity of the group. It follows that a has order 2. Since N is a normal subgroup of
G, by definition, we see that .g −1 ag ∈ G for any .g ∈ G. Since .a /= e, we must have
.g
−1 ag = a. That is, .ag = ga. Hence we have shown that .ag = ga for all .g ∈ G.

Since .ag = ga, one can easily show by the principle of mathematical induction that
.(ag) = a g for any .g ∈ G, and any nonnegative integer n. The quotient group
n n n

.G/N is of order 13, and since any group of prime order is cyclic, this quotient group

is cyclic. Let .b ∈ G\N . Then the coset bN has order 13. In particular the order of
b cannot be 2. Since its order has to divide 26, the order of b must be either 13 or
26. If the order is 26, then G is cyclic with generator b. If the order is 13, then since
.ab = ba, we must have .(ab)
13 = a 13 b13 = a, and therefore .(ab)26 = a 2 = e, and

hence ab has order 26. Therefore ab generates G, and therefore G is cyclic.

S2012-6 Suppose contrary to what is asserted, the indicated sum is a rational

number. Then there are positive integers a and b such that
∞
⎲ 1 a
. = .
(n!)k b
n=0

We multiply both sides by .((b + 1)!)k and rewrite the sum as follows:
∞
⎲ ((b + 1)!)k
a
((b + 1)!)k
. =
b (n!)k
n=0

⎲
b+1 ∞
⎲
((b + 1)!)k ((b + 1)!)k
= + .
(n!)k (n!)k
n=0 n=b+2

Thus
∞
⎲ ⎲
b+1
((b + 1)!)k ka ((b + 1)!)k
.
k
= ((b + 1)!) − . (35.4)
(n!) b (n!)k
n=b+2 n=0

It is clear that the right-hand side is a positive integer. We now proceed to show that
the left-hand side is not an integer, thereby getting the desired contradiction. For
.n ≥ b + 2, we write .n = (b + 2) + j for .j ≥ 0 and hence
126 35 2012 Solutions

⎛ ⎞k
((b + 1)!)k (b + 1)!
. =
(n!)k ((b + 2) + j )!
⎛ ⎞k
1
=
(b + 2 + j ) · · · (b + 2)
⎛ ⎞k
1
≤ .
(b + 2)j +1

As a consequence we see that

∞
⎲ ∞
⎲
((b + 1)!)k ((b + 1)!)k
. =
(n!)k (((b + 2) + j )!)k
n=b+2 j =0
∞ ⎛
⎲ ⎞j +1
1
≤
(b + 2)k
j =0

1
= .
(b + 2)k − 1

1
Note that . < 1, showing that the left-hand side sum in (35.4) cannot
(b + 2)k − 1
be an integer.

S2012-7 Note that

⎰ |1
1 n | n
. nx(1 − x 2 )n dx = − (1 − x 2 )n+1 || = . (35.5)
0 2(n + 1) 0 2(n + 1)

As a result of this, the claim will follow once we show that

⎰ 1
. lim nx(1 − x 2 )n [f (x) − f (0)] dx = 0.
n→∞ 0

To this end, let .ϵ > 0 be given. Since f is continuous at .x = 0, there is,

corresponding to the given .ϵ (but not depending on n), a number .δ ∈ (0, 1) such
that if .0 ≤ x < δ, then .|f (x) − f (0)| < ϵ. First let us note that
⎰ 1 ⎰ δ
. nx(1 − x 2 )n |f (x) − f (0)| dx = nx(1 − x 2 )n |f (x) − f (0)| dx
0 0
⎰ 1
+ nx(1 − x 2 )n |f (x) − f (0)| dx
δ

= In + I In . (35.6)
35 2012 Solutions 127

We estimate each of the summands .In and .I In . For .In , we use the continuity of f at
0. Thus for any n,
⎰ δ ⎰ 1
In =
. nx(1 − x ) |f (x) − f (0)| dx ≤ ϵ
2 n
nx(1 − x 2 )n dx
0 0
ϵn
=
2(n + 1)
< ϵ/2.

Next we estimate .I In . For this, we use the boundedness of f on .[0, 1]: There is
some .M > 0 so that .|f (x)| ≤ M for all .0 ≤ x ≤ 1. For any .n > 0, the function
.g(x) = n(1 − x ) is decreasing on .[0, 1], so on the interval .[δ, 1],
2 n

0 ≤ g(x) ≤ n(1 − δ 2 )n .
.

We estimate:
⎰ 1
.I In = nx(1 − x 2 )n |f (x) − f (0)| dx
δ
⎰ 1
≤ 2M nx(1 − x 2 )n dx
δ
⎰ 1
≤ 2M nx(1 − δ 2 )n dx
δ

< Mn(1 − δ 2 )n .

On noting that . lim n(1 − δ 2 )n = 0, there is some N depending on .δ and .ϵ so that

n→∞
if .n > N then .I In < ϵ/2. Therefore, from (35.6) and the above estimates we find
that for sufficiently large n,
|⎰ | ⎰
| 1 | 1
| nx(1 − x 2 )n (f (x) − f (0)) dx || ≤ nx(1 − x 2 )n |f (x) − f (0)| dx
.
|
0 0
= In + I In ≤ ϵ.

Therefore, we have
⎰ 1
. lim nx(1 − x 2 )n (f (x) − f (0)) dx = 0 (35.7)
n→∞ 0

⎰ 1
as claimed. Finally, adding . lim nx(1 − x 2 )n f (0) dx to both sides of (35.7),
n→∞ 0
and using (35.5), we find that
128 35 2012 Solutions

⎰ 1 ⎰ 1 1
. lim nx(1 − x 2 )n f (x) dx = lim nx(1 − x 2 )n f (0) dx = f (0).
n→∞ 0 n→∞ 0 2

S2012-8 First we rewrite the given double integral as

⎰ ⎰ ⎰ ⎰ ⎰ ⎰
. f (x, y) dx dy = f (x, y) dx dy − f (x, y) dx dy, (35.8)
R D1 D2

where .D1 and .D2 are the closed disks

D1 = {(x, y) : x 2 + (y + 1)2 ≤ 9} and .D2 = {(x, y) : x 2 + (y − 1)2 ≤ 1}.
.

The following property of a function .f (x, y) that is harmonic in an open set

containing a (closed, positive radius) disk .D centered at .(x0 , y0 ) can be interpreted
as a Mean Value property:
⎰ ⎰
1
. f (x, y) dx dy = f (x0 , y0 ). (35.9)
area(D) D

We now apply formula (35.9) to (35.8) to get

⎰ ⎰ ⎰ ⎰ ⎰ ⎰
. f (x, y) dx dy = f (x, y) dx dy − f (x, y) dx dy
R D1 D2

= area(D1 )f (0, −1) − area(D2 )f (0, 1)

= 9πf (0, −1) − πf (0, 1).

To prove the Mean Value Property (35.9), let r be the radius of D, and consider
any disk .E of radius .0 < ρ ≤ r centered at .(x0 , y0 ). We give the curve .∂E a
counterclockwise orientation and apply Green’s Theorem to find
⎰ ⎰
0=
. [fxx (x, y) + fyy (x, y)] dx dy
E
⎰ ⎰
= [(fx )x (x, y) − (−fy )y (x, y)] dx dy. (35.10)
E
⎰
⎡ ⎤
= −fy (x, y) dx + fx (x, y) dy (35.11)
∂E
⎰ 2π (
= ρfy (x0 + ρ cos θ, y0 + ρ sin θ ) sin θ
0
+ ρfx (x0 + ρ cos θ, y0 + ρ sin θ ) cos θ ) dθ.

In the last integral we used the following parametrization for the counterclockwise
oriented circle .∂E.
35 2012 Solutions 129

x = x0 + ρ cos θ,
. y = y0 + ρ sin θ, 0 ≤ θ ≤ 2π.

On dividing both sides of the last equation by .ρ, and observing that the multivariable
chain rule applies, we see that for any .0 < ρ < r,
⎰ 2π
0=
. (fy (x0 + ρ cos θ, y0 + ρ sin θ ) sin θ +fx (x0 + ρ cos θ, y0 + ρ sin θ ) cos θ ) dθ
0
⎰ 2π d
= (f (x0 + ρ cos θ, y0 + ρ sin θ )) dθ. (35.12)
0 dρ

Integrating in polar coordinates, and using the Fundamental Theorem of Calculus,

an interchange of integrals of continuous functions, and (35.12), we obtain
⎰ ⎰
. f (x, y) dx dy
D
⎰ 2π ⎰ r
= f (x0 + R cos θ, y0 + R sin θ )R dR dθ
0 0
⎰ 2π ⎰ r ⎡ ⎰ R ⎤
d
= f (x0 , y0 ) + (f (x0 + ρ cos θ, y0 + ρ sin θ )) dρ R dR dθ
0 0 ρ=0 dρ
⎰ 2π ⎰ r
= f (x0 , y0 )R dR dθ
0 0
⎰ r ⎰ R ⎰ 2π d
+ (f (x0 + ρ cos θ, y0 + ρ sin θ )) R dθ dρ dR
0 ρ=0 0 dρ
= area(D) · f (x0 , y0 ) + 0.

This proves the claim (35.9).

Remark from the Editors It is enough to assume f is harmonic on an open set

containing .D1 , but not enough to assume only that f is harmonic on some open
set containing .R. The problem statement’s hypothesis that f is twice continuously
differentiable is used only to justify the application of Green’s Theorem in the step
equating line (35.10) with line (35.11). There are statements of Green’s Theorem
with weaker hypotheses (see [CZ]) that could be applied instead; for example, the
step from (35.10) to (35.11) still follows from assuming only that .fxx + fyy is
continuous and .(fx , fy ) is differentiable as a function .R2 → R2 .
Chapter 36
2013 Solutions

The 48th competition was held at Indiana University East for the first time.
The 2013 problem set was prepared by a faculty committee at the host institution.
The 2013 problem statements begin on page 29 in Chap. 13.

The 2013 competition’s first place team from Valparaiso. Left to Right: Michael Stuck, Julia Yuan,
and Timothy Goodrich

S2013-1 We show that the sequence is monotone increasing and bounded above.
It then follows from the Monotone Convergence Theorem that the sequence
converges. We show that for every .n ∈ N, we have .sn < 2 and .sn+1 ≥ sn . This

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 131
A. Coffman et al., The Indiana College Mathematics Competition (2001–2023),
Problem Books in Mathematics, https://doi.org/10.1007/978-3-031-62768-2_36
132 36 2013 Solutions

√
is done by induction. For .n = 1, we have .s1 = 1 < 2 and .s2 = 2 > 1 = s1 .
Assume that√for some .n ∈ N, we have √
√ .sn < 2 and .sn+1 > sn . Then .sn+1 =
√
1 + sn < 1 + 2 < 2. Also, .sn+2 = 1 + sn+1 > 1 + sn = sn+1 . By the
principle of Mathematical Induction, we obtain that for every .n ∈ N, we have
.sn < 2 and .sn+1 < sn . Because the sequence is bounded above and increasing,

it converges to some positive real number. Let .s = lim sn . We observe that

√ √ n→∞
.s = lim sn = lim sn+1 = lim 1 + sn = 1 + s. Solving the equation
n→∞ n→∞ n→∞√
√ 1+ 5
.s = 1 + s for .s > 0 yields .s = .
2
S2013-2

(a) We let .C0 ∈ C be an arbitrary element of the collection. Because .C0⋂

is compact,
it is closed and bounded, by the Heine–Borel Theorem. Also, .K = C ⊆ C0 ,
C∈C
so K is bounded. Each member C of the ⋂family .C is compact and therefore
closed. Therefore, the intersection .K = C is closed. Because K is closed
C∈C
and bounded, it is compact by the Heine–Borel Theorem. ⨆
⨅
(b) Let .Cn = [−n, n] for all .n ∈ N. Each set .Cn is closed
⋃ and bounded and is
therefore compact. Let .C = {Cn : n ∈ N}. Then . C = R, which is not
C∈C
compact (it is not bounded).

S2013-3

(a) We distinguish two cases.

• If .m > n, there are no one-to-one functions from A to B by the pigeonhole
principle.
• If .m ≤ n, then the principle of multiplication shows there are nP m or
.P (n, m) =
(n−m)! = n · (n − 1) · · · (n − m + 1) one-to-one functions from
n!

A to B. Equivalently, if one enumerates the elements in A as .(ai )m i=1 and the

elements of B as .(bj )nj=1 , there are clearly n available images for domain
element .a1 ; without loss of generality, say this element is .b1 . If one is to
maintain the one-to-one property, there are now .n − 1 available images for
.a2 ; and without loss of generality, say the image is .b2 . This pattern continues,

yielding .n − 2 possible images for .a3 , .n − 3 images for .a4 , and so on. This
also yields the answer .n · (n − 1) · · · (n − m + 1).
(b) If m is not equal to n, there is no one-to-one and onto function from A to B. If
m is equal to n, then by the principle of multiplication there are .n! one-to-one
and onto functions from A to B.
36 2013 Solutions 133

S2013-4 An element .a ∈ Zpq is a generator of .Zpq if and only if a and pq are

relatively prime. Because p and q are primes, the elements of .Zpq that are not
relatively prime to pq are the multiples of p and the multiples of q. The multiples
of p are .p, 2p, . . . , (q − 1)p (i.e., .q − 1 multiples). Using a similar argument, we
see that there are .p − 1 multiples of q. Also, 0 is not a generator. Therefore, there
are .(q − 1) + (p − 1) + 1 = p + q − 1 elements of .Zpq that are not generators,
leaving .pq − p − q + 1 elements that are generators.

S2013-5 Because the index of H in G is 2, there are exactly two left cosets of H
in G, and there are exactly two right cosets. The left cosets are H itself and a coset
of the form aH , for some .a ∈ G. Likewise, the right cosets are H and a set of the
form H b for some .b ∈ G. Observe that .H ∩ aH = Ø and .H ∪ aH = G; likewise
.H ∩ H b = Ø and .H ∪ H b = G. It follows that .aH = H b, which means that the

left and the right cosets of H coincide, making H a normal subgroup of G.

S2013-6A

n 123456 7
(a) .
fn 1 1 2 3 5 8 13
(b) For .n = 1, we observe .f2(1)+1 = f3 = 2 and .f1+12 + f 2 = (1)2 + (1)2 = 2.
1
For .n = 2, we observe .f2(2)+1 = f5 = 5 and .f2+12 + f 2 = (2)2 + (1)2 = 5.
2
For .n = 3, we observe .f2(3)+1 = f7 = 13 and .f3+1
2 + f 2 = (3)2 + (2)2 = 13.
3
(c) For .n = 1 and .n = 2, the formula has been verified in part (b). Therefore, the
basis steps hold for mathematical induction. Now assume, for the strong form
of mathematical induction, the identity holds for all values of n up to .n = k − 1.
Then

f2k−3 = fk−1
.
2
+ fk−2
2

and

f2k−1 = fk2 + fk−1

.
2
.

Now we need to verify that the identity holds for .n = k. In order to do this, we
calculate .f2k+1 .

f2k+1 = f2k + f2k−1

= f2k−1 + f2k−2 + f2k−1
.
= 2f2k−1 + (f2k−1 − f2k−3 )
= 3f2k−1 − f2k−3 .

Substituting the induction hypothesis, we can write the last expression as

134 36 2013 Solutions

( )
f2k+1 = 3 fk2 + fk−1
2 − fk−1
2 − f2
k−2
= 3fk2 + 2fk−1
2 − (f − f
k k−1 )
2

= 2fk + fk−1 + 2fk fk−1

2 2

. = 2fk2 + (fk+1 − fk )2 + 2fk (fk+1 − fk )

= 2fk2 + (fk+1 − fk ) (fk+1 − fk + 2fk )
= 2fk2 + (fk+1 − fk ) (fk+1 + fk )
= fk+1
2 + f 2.
k

This completes the induction step.

S2013-6B An alternate
⎾ ⏋approach to (c)
⎾ is to recall⏋ (from P2003-6 in Chap. 3) that
11 fn+1 fn
if Q is the matrix . , then .Q =
n . From
10 fn fn−1
⎾ ⏋
f2n+1 f2n
.Q =
2n
f2n f2n−1
⎾ ⏋ ⎾ ⏋
fn+1 fn f f
= Qn · Qn = · n+1 n
fn fn−1 fn fn−1
⎾ ⏋
2 + f2
fn+1 fn+1 fn + fn fn−1
= n ,
fn+1 fn + fn fn−1 fn2 + fn−1
2

comparing the upper left entries establishes the claim.

S2013-7 Since all these numbers are positive, it is sufficient to prove

4mM 4ab
. ≤ .
(m + M) 2
(a + b)2

This is equivalent to proving

m
4M 4 ab
. ( ) ≤ ( )2 .
m 2
1+ M 1 + ab

4x
Consider the function .f (x) = . This function is increasing on .[0, 1]. To see
⎛ + x)2
(1 ⎞
d 4x 1−x
this, we note that . =4 exists for .x > −1 and is positive
dx (1 + x) 2
(x + 1)3
for .0 < x < 1.
36 2013 Solutions 135

m a
From .0 < m ≤ a and .M > 0, we obtain .0 < ≤ . Because .b ≤ M, we also
M M
a a a m a
have . ≤ . Finally, .a ≤ b, so . ≤ 1. Therefore, .0 < ≤ ≤ 1.
M b b M b
Hence we have
⎛m⎞ 4m ⎛a ⎞ 4a
f
. = ( M )2 ≤ f = ( b )2 .
M 1+ M
m b 1 + ab

2013-8 We let .F5 denote the number of pentagons and .F6 denote the number of
hexagons. We consider the soccer ball to be a convex polyhedron, with .F = F5 +F6
faces, E edges, and V vertices. By Euler’s Formula .V − E + F = 2.
Each vertex is of valence 3. We may think of placing an observer on each vertex, and
let the observers report the number of faces they see. Each observer reports seeing
three faces. Each face is observed by as many observers as there are corners on the
face, so we obtain .3V = 5F5 + 6F6 .
Now place an observer in every hexagon, and let them report the number of
pentagons that border their hexagon. There are .F6 observers, each reporting three
pentagons, for a total of .3F6 reports. Each pentagon is bordered by five hexagons,
so each pentagon is reported by five different observers, and .3F6 = 5F5 .
Now place an observer into each of the faces, and let them report the number of
edges they see. Each edge will be observed by two observers, so .2E = 5F5 + 6F6 .
Beginning with the equation .V − E + F = 2 (and multiplying by 6), we obtain
.6V − 6E + 6F = 12. Substituting .3V = 5F5 + 6F6 and .2E = 5F5 + 6F6 and
.F = F5 + F6 , we obtain .2(5F5 + 6F6 ) − 3(5F5 + 6F6 ) + 6(F5 + F6 ) = 12, or

.F5 = 12.

From the equation .3F6 = 5F5 , we get .F6 = 20. Therefore, the soccer ball has 12
black pentagons and 20 white hexagons.

Remark from the Editors Note that the .3F6 = 5F5 equation was not used until
the last paragraph, so more generally, any convex polyhedron with pentagon and
hexagon faces and valence three vertices must have exactly 12 pentagons. In the
2013 competition, one team’s submitted answer was a well-drawn picture of a
semitransparent soccer ball exhibiting 12 pentagons and 20 hexagons. The grader
assigned partial credit for showing existence but not uniqueness.
Chapter 37
2014 Solutions

The .49th competition was held at Indiana University—Purdue University Fort

Wayne, coinciding with the celebration of the .50th anniversary of the founding of
that campus.
The 2014 problem set was prepared by Professors Dan Coroian and Marc Lipman
at the host institution.
The 2014 problem statements begin on page 31 in Chap. 14.

A tied-for-first place team from Taylor University, with the IPFW Mastodon sculpture: L to R:
Ethan Gegner, Josh Keirs, Claire Spychalla

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 137
A. Coffman et al., The Indiana College Mathematics Competition (2001–2023),
Problem Books in Mathematics, https://doi.org/10.1007/978-3-031-62768-2_37
138 37 2014 Solutions

S2014-1 The expression (14.1) simplifies to

a 3/2 x 1/3 (x 1/6 − a 1/6 )

f (x) =
. (37.1)
a 1/4 (a 3/4 − x 3/4 )

and is continuous on .[0, a) ∪ (a, ∞), with the .x → 0+ limit equal to 0.

For the .x → a limit, the . 00 form of L’Hôpital’s Rule applies.
√ √
3
a3x − a a2x a 3/2 x 1/2 − a 5/3 x 1/3
. lim √ = lim
x→a 4
a − ax 3 x→a a − a 1/4 x 3/4
a 3/2 12 x −1/2 − a 5/3 13 x −2/3
(LH R) = lim
x→a 0 − a 1/4 34 x −1/4
1
2a − 13 a
=
− 34
2a
=− .
9
For the .x → +∞ limit, LHR could be used again, but it is easier to notice that f
satisfies .|f (x)| < Cx −1/4 for large x, so there is a horizontal asymptote .f (x) → 0
as .x → ∞. From (37.1), .f (x) < 0 for all .x ∈ (0, a) ∪ (a, ∞), so the maximum
value is .f (0) = 0.

Remark from the problem authors Finding critical points to locate maximum values
can be attempted, but this is a difficult calculation and a waste of time. The
formula (14.1) is a variation on an expression appearing in an early calculus book
by L’Hôpital (and Johann Bernoulli) [S].
k k−1 k−1
S2014-2 For integer .k ≥ 1, .f (x (2 ) ) = f ((x (2 ) )2 ) = f (x (2 ) ), so by
k
induction, .f (x (2 ) ) = f (x) for all integers .k ≥ 0. Given .ε > 0, there is some
.δ ∈ (0, 1) and some .N ∈ (1, ∞) so that if .0 < t < δ or .t > N, then

.|f (t) − f (1)| < ε. If .0 < x < 1, then there is some integer k such that

.k > log2 (ln(δ)/ ln(x)), which is equivalent to .0 < x

(2k ) < δ, and if .x > 1, then

there is some integer k such that .k > log2 (ln(N )/ ln(x)), which is equivalent to
.N < x
(2k ) , so in either case, .|f (x) − f (1)| = |f (x (2k ) ) − f (1)| < ε. Since .ε was

arbitrary, .f (x) = f (1).

Remark from the problem authors More informal limit arguments can be attempted,
but at the risk of taking some unjustified steps.
opp
S2014-3A From .sin = hyp , .sin x = a/d, .sin y = b/d, and .sin z = c/d, where a,
b, c are the side lengths of the box and d is the long diagonal length.
37 2014 Solutions 139

⎡ ⎤
abc
1
. det(A) = det ⎣ c a b ⎦ .
d3
bca

The absolute value of

⎡ ⎤
abc
. det ⎣ c a b ⎦

bca

is the volume of a parallelepiped with side lengths all equal to d. By the scalar triple
product formula, such a volume is maximized when the parallelepiped has all right
angles, so it is a cube with volume .d 3 . The claimed inequality follows.

opp
S2014-3B From .sin = hyp , .sin x = a/d, .sin y = b/d, and .sin z = c/d, where
a, b, c are the side lengths of the box and d is the long diagonal length. By the
Pythagorean Theorem,

a 2 + b2 + c2
. sin2 (x) + sin2 (y) + sin2 (z) = = 1.
d2
This gives the diagonal entries in the matrix product:
⎡ ⎤
1s s
.AA = ⎣ s 1 s ⎦,
T

s s 1

where the other entries are:

bc + ac + ab bc + ac + ab
s = sin(z) sin(y)+sin(x) sin(z)+sin(y) sin(x) =
. = 2 .
d2 a + b2 + c2
140 37 2014 Solutions

By the arithmetic-geometric mean inequality, .bc + ac + ab ≤ 12 (b2 + c2 ) + 12 (a 2 +

c2 ) + 12 (a 2 + b2 ), so .0 ≤ s ≤ 1. On the interval .0 ≤ s ≤ 1,

f (s) = det(AAT ) = det(A)2 = 1 − 3s 2 + 2s 3

is decreasing (with .f ' (s) = 6s(s − 1) ≤ 0) from maximum value 1 to minimum

value 0, and the claimed inequality .|det(A)| ≤ 1 follows.
opp
S2014-3C From .sin = hyp , .sin x = a/d, .sin y = b/d, and .sin z = c/d, where a,
b, c are the side lengths of the box and d is the long diagonal length. Expanding the
determinant gives
⎡ ⎤
abc
⎣ c a b ⎦ = a + b + c − 3abc .
1 3 3 3
. det(A) = det
d3 d3
bca

The RHS is nonnegative by the three-variable arithmetic-geometric mean inequality,

3 (a + b + c ) ≥ (a b c )
.
1 3 3 3 3 3 3 1/3 . Using .sin x = a/d ≤ 1, and similarly .b/d ≤ 1,

.c/d ≤ 1, the RHS also has this upper bound:

a 3 + b3 + c3 − 3abc a a2 b b2 c c2 3abc
.
3
= 2
+ 2
+ − 3
d dd dd d d2 d
a 2 + b2 + c2 3abc
≤ 2
− 3
d d
3abc
= 1 − 3 ≤ 1.
d

So we get a stronger conclusion, .0 ≤ det(A) ≤ 1 − 3 sin(x) sin(y) sin(z) ≤ 1.

⎰x
S2014-4 Let .F (x) be the function .2x − 1 − 0 f (t)dt, which is continuous on
⎰1 ⎰1
.[0, 1] and satisfies .F (0) = −1 and .F (1) = 1 −
0 f (t)dt ≥ 1 − 0 1dt = 0. By
the Intermediate Value Theorem, .F (x) = 0 has at least one solution .x ∈ [0, 1]. This
solution is unique because F is increasing on .[0, 1]: For .0 ≤ a < b ≤ 1,
⎰ b
F (b) − F (a) = 2(b − a) −
. f (t)dt ≥ 2(b − a) − 1(b − a) = b − a > 0.
a

Remark from the problem authors If .f (t) were continuous, then F could be proved
increasing using the Fundamental Theorem of Calculus: .F ' (x) = 2 − f (x) ≥ 1.
However, the problem specifically omits this hypothesis.

S2014-5 Let .AB = b, .BC = h, .AM = y, .MB = x, let .θ be half the angle ABC,
and let .α be the angle BMC. By the Law of Sines,
37 2014 Solutions 141

√
sin θ sin(π − α) sin α sin θ sin α h AC b 2 + h2
. = , = =⇒ = = = .
y AC h x sin θ x y y

We have the following system of polynomial equations:

x+y = b
.

1
xh = s
2
bh = s + t
x 2 (b2 + h2 ) = h2 y 2 .

Eliminating y first gives

x 2 (b2 + h2 ) = h2 (b − x)2 =⇒ x 2 b = h2 b − 2h2 x.

Multiplying both sides by .h3 gives

.x 2 bh3 = h4 (hb − 2hx)

(2s)2 (s + t) = h4 (s + t − 2(2s))
⎛ 2 ⎞1/4
4s (s + t)
h=
t − 3s
s+t (s + t)3/4 (t − 3s)1/4
b= = √ .
h 2s

The .b/ h ratio can be computed directly for .t = 4s, or as

√
b bh s + 4s 5s 5
. = 2 =⎛ ⎞1/2
=√ = .
h h 4s 2 (s+4s) 20s 2 2
4s−3s

Remark from the problem authors The equality of ratios . xh = AC

y from the first step
is also known as the “bisector theorem” for triangles.

S2014-6 There are eight types of students with the following populations:

. # RH BL boy = 2
# RH BR boy = 9
# LH BR girl = 5
142 37 2014 Solutions

# LH BR boy = 13 − 5 = 8
# RH BR girl = 26 − 9 − 13 = 4
# LH BL girl = x
# RH BL girl = y
# LH BL boy = z.

From equal numbers of boys and girls, .x+y+9 = 19+z. From equal numbers of LH
and RH, .x + z + 13 = y + 15. From one fourth LH girls, .4(x + 5) = x + y + z + 28.
This is a system of three linear equations in three unknowns. Standard solution
methods give the unique answer .x = 6, .y = 7, and .z = 3, so the total population is
.x + y + z + 28 = 44, with seven RH BL girls.

Remark from the problem authors Drawing a Venn diagram may be helpful.

S2014-7 For the first part,

F (e) = f (e) ∗ f (a · e) ∗ f (a 2 · e) ∗ . . . ∗ f (a n−1 · e)

= f (e) ∗ f (a) ∗ f (a 2 ) ∗ . . . ∗ f (a n−1 ),

F (a) = f (a) ∗ f (a 2 ) ∗ f (a 2 · a) ∗ . . . ∗ f (a n−2 · a) ∗ f (a n−1 · a)
= f (a) ∗ f (a 2 ) ∗ f (a 3 ) ∗ . . . ∗ f (a n−1 ) ∗ f (e) = F (e),

using the property that .f (e) commutes with other .f (g) at the last step. By the
assumption that .a /= e, F is not one-to-one.
For the second part, there are lots of examples. (A correct answer must have explicit
examples of G, n, a, and f .) A simple one is to let G be a two-element group .{e, a},
so .n = 2, and to define .f : G → S3 by .f (e) = (12) and .f (a) = (23), or any other
pair of non-commuting elements in .S3 . Then .F (e) = f (e) ∗ f (a) = (12) ◦ (23) =
(123) and .F (a) = f (a) ∗ f (e) = (23) ◦ (12) = (132), so F is one-to-one.

S2014-8 Let x be the number; the cube roots suggest that x may be a root of a
cubic polynomial, so try expanding .x 3 . The following calculation has convenient
cancelations and shows x satisfies .x 3 = 12 + 5x:
/ ⎛ / ⎞1/3
847 847 847
x3 = 6 +
. + 3 (6 + )(36 − )
27 27 27
⎛ / ⎞1/3 /
847 847 847
+3 (36 − )(6 − ) +6−
27 27 27
37 2014 Solutions 143

⎛ / / ⎞
847 1/3 847 1/3 847 1/3
= 12 + 3(36 − ) (6 + ) + (6 − )
27 27 27
⎛ ⎞1/3
125
= 12 + 3 x = 12 + 5x.
27

The only real root of .x 3 − 5x − 12 = (x − 3)(x 2 + 3x + 4) is .x = 3; the conclusion

is that x is rational.

Remark from the Editors This problem is similar to P2009-4 and to ICMC P1969-3
[AFMC].
Chapter 38
2015 Solutions

The 2015 competition was held at Taylor University. On the occasion of the .50th
annual ICMC event, the Indiana Section of the MAA provided themed T-shirts to
the contest participants. 2015 was also the Centennial year of the MAA.
The 2015 problem set was prepared by Professor Brian Rice at Huntington
University and Professor Feng Tian at Trine University.
The 2015 problem statements begin on page 35 in Chap. 15.

The team from Indiana—Purdue Fort Wayne, third place in 2015: L to R: Soﬁa Sorokina, Altun
Shukurlu, Vreneli Brenneman

S2015-1

(a) The number 2015 has the super-3 representation .2015 = 37 − 35 + 34 − 32 − 30 .

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 145
A. Coffman et al., The Indiana College Mathematics Competition (2001–2023),
Problem Books in Mathematics, https://doi.org/10.1007/978-3-031-62768-2_38
146 38 2015 Solutions

(b) Without loss of generality we may let n be a nonnegative integer, since given
a super-3 representation of n, we can ﬁnd a super-3 representation of .−n by
changing the signs of all the .ak . Now, we prove this by strong induction on n.
As given in the problem statement, 0 has a super-3 representation; this is our
base case. Now, let .n > 0 and suppose that every .0 ≤ x < n has a super-3
representation. Then n is either a multiple of 3, one more than a multiple of 3,
or one less than a multiple of 3. Let q be the nearest multiple of 3 to n, and
let .r = q/3. Certainly .0 ≤ r < n, so by the inductive step it has a super-3
⎲
m ⎲
m
representation; say .r = ak · 3pk . Then .q = ak · 3pk +1 ; then either .n = q,
k=1 k=1
and that sum is a representation of n, or .n = q ± 1. In the latter case:
⎲
m
• If .n = q + 1, then .n = ak · 3pk +1 + 30 (so .am+1 = 1 and .pm+1 = 0).
k=1
⎲
m
• If .n = q − 1, then .n = ak · 3pk +1 − 30 (so .am+1 = −1 and .pm+1 = 0).
k=1
In any case this gives a super-3 representation of n since the exponents of 3 are
all distinct. Each positive exponent is one more than its corresponding exponent
in a known super-3 representation (thus all positive exponents are distinct), and
if one of the exponents is 0, it is the only such exponent. This proves that n has
a super-3 representation, completing the inductive step and the proof.

Remark from the problem author Several other solutions are possible, including
different induction arguments and a construction based on base-3 representations.

S2015-2A Rewrite the left-hand side as

⎛⎛ ⎞ n−k ⎛ ⎞ ⎞
⎲
n
n k⎲ n−k i
. 2 2 .
k i
k=0 i=0

Applying the binomial theorem to the interior sum, we obtain .(2 + 1)n−k = 3n−k ,
so we ﬁnd that the left-hand side is equal to
n ⎛ ⎞
⎲ n
. 2k 3n−k .
k
k=0

Applying the binomial theorem again, we ﬁnd that the left-hand side is equal to
(2 + 3)n = 5n as required.
.

S2015-2B The term inside the double sum counts the number of ways to choose,
from a group of n people, two disjoint committees A and B of sizes k and i,
38 2015 Solutions 147

respectively, and then to choose a (possibly empty) third committee C from among
the members of both committees A and B. The double sum then adds all these up
over all possible values of k and i. Hence the sum as a whole counts the number of
ways to choose from n people two disjoint committees A and B, and then a third
committee C from among the members of A and B. We can count this number
in a different way: Each person can be either (1) on no committees at all, (2)
on committee A only, (3) on committee B only, (4) on committees A and C, or
(5) on committees B and C. Since each person independently has ﬁve possible
assignments, the total number of ways to do this is .5n , as required.

S2015-3 Since .AB and .CD intersect, the quadrilateral ACBD is planar and
inscribed in the circle that its plane cuts from the sphere. By the power of a point
theorem, .AX · BX = CX · DX, whence .1 · 4 = 2 · DX and .DX = 2. Since .AB and
.CD are perpendicular, the area N of ACBD is .N = AB · CD = (5 · 4) = 10.
1 1
2 2
The same argument shows that .1 · 4 = AX · BX = EX · F X = 3 · F X, and thus
.F X = 43 . Since .EF is perpendicular to the plane containing ACBD, the octahedron
is divided by that plane into two right pyramids with base area .N = 10 and heights
.EX = 3 and .F X = , respectively. These two pyramids have volumes . EX · N =
4 1
3 3
1
3 (3 · 10) = 10 and . F X · N = ( · 10) =
1
3
1 4
3 3
40
9 . Hence the volume of the octahedron
40 130
is .10 + = .
9 9
Remark from the problem author The argument above can be used to prove the
little-known fact that the power of a point theorem is true in three dimensions as
well as in two.

Remark from the Editors At the 2015 ICMC event, at least one team’s submitted
solution included a veriﬁcation of the pyramid volume formula by calculating an
integral.

S2015-4

(a) The minimum number of trials is 3. In the ﬁrst trial, turn on switches 1, 2, 3, and
4; in the second, switches 1, 2, 5, and 6, and in the third, 1, 3, 5, and 7. Then the
light that was on in all three trials is that controlled by switch 1, that which is on
in trials 1 and 2, but not trial 3, is controlled by switch 2, and so on. It is easy to
verify that each light is on in a difference subset of the trials, so this determines
which switches control which lights.
(b) Observe that if two switches x and y are always on or off together in our
trials, then our trials cannot distinguish which of them control which light.
For instance, if in every trial, switches 1 and 2 are either both on or both off,
and these switches control lights A and B, we will not be able to determine
whether switch 1 controls light A or light B. Therefore, if we can use a set
of trials to determine which switches control which lights, it must be the
148 38 2015 Solutions

case that the function from switches to subsets of tests given by .f (x) =
{α | Light x is on in Test α} is one-to-one. Since the number of subsets of all
trials is 2 to the power of the number of trials, and the number of switches is 8,
it follows that there must be at least three trials.

Remark from the problem author The same argument can be used to show that in
the general case of n switches, we need to use at least .⎾log2 (n)⏋ trials to determine
which switches control which lights.

S2015-5

(a) Let .a1 , a2 , . . . be an infinite sequence of zeros of f in the interval .[0, 1]. Since
every bounded infinite sequence has a convergent subsequence, we may find
such a convergent subsequence .b1 , b2 , . . .. Let .x = lim bn . Certainly, .x ∈
n→∞
[0, 1]. Now, there are either infinitely many .bn less than x or infinitely many .bn
greater than x; without loss of generality, suppose the former. It then follows
that we may choose an infinite increasing sequence .c1 , c2 , . . . from among the
.bn ; simply begin with .c1 < x, and continue choosing .cn+1 from among the

.bi , such that .cn < cn+1 < x. (We can always do this because otherwise we

would have infinitely many .bi which are less than .cn , which is impossible since
. lim bn = x > cn .) Since this increasing sequence is a subsequence of the .bn ,
n→∞
it necessarily converges to x.
Since f is differentiable, it is continuous, and it follows that .f (x) =
f ( lim cn ) = lim f (cn ) = 0. Now by Rolle’s Theorem, there is a zero
n→∞ n→∞
of .f ' between .cn and .cn+1 for every .n ≥ 1; let .cn be a zero of .f ' between .cn
(1)
(1) (1)
and .cn+1 . Then the sequence .c1 , c2 , . . . is an increasing sequence of zeros of
' '
.f , and by the squeeze theorem it has limit x. As .f is itself differentiable and
'
thus continuous, it follows as above that .f (x) = 0. Continuing by induction,
we find that for every n, there is an increasing sequence .c1(n) , c2(n) , . . . of zeros
of .f (n) with limit x, and hence that .f (n) (x) = 0.
This shows that our chosen x satisfies the required condition.
(b) One such function is given by
⎧ ⎛ ⎞ ⎛ ⎞
⎨sin 1 exp − 1 x /= 0
.f (x) =
x x2
⎩0 x = 0.

Remark from the problem author The function given above for the second part
is a modiﬁcation of a classic example showing that a nonzero function can have
identically zero Taylor series at 0.
38 2015 Solutions 149

Remark from the Editors The differentiability of the function .exp(−1/x 2 ) is

considered in ICMC P1982-4 [AFMC].

S2015-6 Let G be a k-color-planar graph. Each of the k sets of edges with the same
color forms a planar graph on the vertices of G. By the Four Color Theorem, each
of these graphs can be 4-colored. Doing so, each vertex of G gets a k-tuple of colors
corresponding to its color in each of the k planar graphs. If we assign a different
color to each k-tuple, this gives us a .4k -coloring of G (since adjacent vertices must
differ in at least one coordinate as they are adjacent in some one of the k planar
graphs). Thus, every k-color-planar graph can be .4k -colored (so .ck = 4k works).

Remark from the problem author Is this value .ck = 4k best possible? We do not
know and would be interested in a proof either way.

S2015-7

(a) Because .∗ is associative, we may omit parentheses without any ambiguity. Let
.a, b ∈ S be arbitrary. We have

.a ∗ b = a ∗ (a ∗ a ∗ b ∗ a)
= a ∗ a ∗ a ∗ (a ∗ a ∗ b ∗ a) ∗ a
= a ∗ (a ∗ a ∗ a ∗ a) ∗ b ∗ a ∗ a
= a∗a∗b∗a∗a
= b ∗ a,

where the assumption that .a ∗a ∗b∗a = b is used in several steps (including one
step where .a ∗ a ∗ a ∗ a = a, the .b = a case). This shows that .∗ is commutative.
(b) It follows that, for any .a, b ∈ S, .b ∗ (a ∗ a ∗ a) = (a ∗ a ∗ a) ∗ b = b, so
.(a ∗ a ∗ a) acts as an identity for .∗. The identity is unique, since if e and f

are both identities, then .e = e ∗ f = f . Call the identity e. Furthermore, since

.a ∗ a ∗ a = e for every .a ∈ S, it follows that .a ∗ a is an inverse for a. Thus,

.∗ is an associative binary operation with an identity and inverses, and therefore

.(S, ∗) is a group.

Moreover, the fact that .a ∗ a ∗ a = e for every .a ∈ S implies that the order of
every element of S divides 3. It follows from Cauchy’s Theorem that .|S| cannot
be divisible by any prime other than 3, so .|S| is a power of 3. (Recall Cauchy’s
Theorem states that if G is a ﬁnite group and .|G| is a multiple of a prime number
p, then there exists an element of G with order p.)

Remark from the problem author In the last step, the fundamental theorem of ﬁnite
abelian groups can be used in place of Cauchy’s theorem.
150 38 2015 Solutions

3
S2015-8 Deﬁne .g(x) = f (x) − x 2 . Then .g(0) = 0 and .g(1) = 0, so by the
Mean Value Theorem, there is some .c ∈ (0, 1) such that .g ' (c) = 0 and hence
' 3√ 9
.f (c) = c. Thus .(f ' (c))2 = c. Applying the Mean Value Theorem to .(f ' )2 on
2 4
the interval .[0, c], we obtain that there is some .a ∈ (0, c) (and hence in .(0, 1)) such
9 9
that .2f ' (a)f '' (a) = ((f ' )2 )' (a) = . Thus .f ' (a)f '' (a) = as required.
4 8
Chapter 39
2016 Solutions

The 51st competition was held at Franklin College. 2016 was also a Centennial
year—the preliminary organization of the Indiana Section of the MAA occurred
at the October 26–28, 1916, meeting of the Mathematics Section of the Indiana
State Teachers’ Association. However, with a delay possibly related to global events
including the war and a pandemic, the ﬁrst meeting of the MAA Indiana Section did
not occur until October 16, 1924 [Da].
The 2016 problem set was prepared by Professors Paul Fonstad, Justin Gash, and
Stacy Hoehn at the host institution.
The 2016 problem statements begin on page 39 in Chap. 16.
S2016-1 To prove that the set is inﬁnite, we will use induction to show that for
.a ∈ A, .n(a)|a whenever a consists of .3m ones (i.e., .n(a) = 3m ) for any integer
.m ≥ 0. For the base case when .m = 0, note that if .n(a) = 3 = 1, .a = 1, so .n(a)|a.
0

Now assume that for some integer .k ≥ 0, .n(b) divides b when b is the element of
A consisting of .3k ones, and consider the element c of A consisting of .3k+1 ones.
k k
Note that .b · 103 consists of .3k ones followed by .3k zeros, while .b · 102·3 consists
of .3k ones followed by .2 · 3k zeros. Therefore, c can be built from three copies of b
via the formula
k k
c = b + b · 103 + b · 102·3
.
k k
= b(1 + 103 + 102·3 ).

Note that .3k = n(b) divides b by the inductive hypothesis, and 3 divides .d = (1 +
k k
103 + 102·3 ) since the sum of the digits of d is 3. Therefore, .3k · 3 = 3k+1 = n(c)
divides c, which completes the proof.
S2016-2A f must be the identity function. In other words, .f (n) = n for all .n ∈ Z.
Step 1. We will use strong induction to show that .f (n) = n for all integers .n ≥ 2.
For the base case, note that .f (2) = 2 is property (i) of f . Now assume that

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 151
A. Coffman et al., The Indiana College Mathematics Competition (2001–2023),
Problem Books in Mathematics, https://doi.org/10.1007/978-3-031-62768-2_39
152 39 2016 Solutions

for some integer .k ≥ 2, .f (n) = n for all integers .2 ≤ n ≤ k, and consider

.f (k + 1).
• Case 1: If .k + 1 is composite, then .k + 1 can be written as the product of two
integers m and n where .2 ≤ m ≤ k and .2 ≤ n ≤ k. Then

f (k + 1) = f (mn)
.

= f (m)f (n) by property (ii) of f

= mn by the inductive hypothesis
= k + 1.

• Case 2: If .k + 1 is prime, then .k + 2 must be composite since .k ≥ 2, so .k + 2

can be written as the product of two integers m and n, where .2 ≤ m ≤ k + 1
and .2 ≤ n ≤ k + 1. We can actually lower the upper bound on both m and
n from .k + 1 to k since . k+2
k+1 < 2 when .k ≥ 2. Then similarly to above,
.f (k + 2) = f (mn) = f (m)f (n) = mn = k + 2. Since .f (k) = k (by the

inductive hypothesis) and .f (k+2) = k+2, we then know that .f (k+1) = k+1
by property (iii) of f .
So far we have shown that .f (n) = n for all integers .n ≥ 2.
Step 2. Note that

f (0) = f (2 · 0) = f (2)f (0) = 2f (0),

so it must be the case that .f (0) = 0. Since .f (0) = 0 and .f (2) = 2, we know
that .f (1) = 1 by property (iii) of f .
Step 3. .f (−1) = −1 because .f (−1) < f (0) = 0 and

(f (−1))2 = f (−1)f (−1) = f ((−1) · (−1)) = f (1) = 1.

To complete the proof, we need to show that .f (n) = n for negative integers
.n ≤ −2. Note that if n is a negative integer, then .−n is a positive integer, so we
know that .f (−n) = −n by our previous work. Then

f (n) = f ((−1) · (−n)) = f (−1) · f (−n) = (−1) · (−n) = n.

Thus, for all integers n, .f (n) = n, so f must be the identity function.

S2016-2B The abovementioned Step 2 could instead be established by using

property (ii) to note that .f (0) · f (0) = f (0 · 0) = f (0) and .f (1) · f (1) =
f (1 · 1) = f (1), so .f (0) = 0 or 1 and .f (1) = 0 or 1. Then by property (iii),
.f (0) = 0 < f (1) = 1.

Then, the abovementioned Step 1 can be given a different proof by induction.

k
Suppose .f (n) = n for .1 ≤ n ≤ 22 . The .k = 0 case is the previously established
39 2016 Solutions 153

(k+1) k k
f (1) = 1 and property (i), .f (2) = 2. For .k > 0, .f (22
. ) = f (22 · 22 ) =
k k k k (k+1)
f (2 ) · f (2 ) = 2 · 2 = 2
2 2 2 2 2 . Because f is strictly increasing, the only
k (k+1)
possible f on the interval from .22 + 1 to .22 − 1 is the identity function.
S2016-3 One such solution would be the points

.(0, 0, 0) .(1, 0, 0)

.(cos(2π/5), sin(2π/5), 0) .(cos(4π/5), sin(4π/5), 0)

.(cos(6π/5), sin(6π/5), 0) .(cos(8π/5), sin(8π/5), 0)

.(0, 0, 1) .(0, 0, −1).

Geometrically, these points represent the origin, the ﬁve vertices of a regular
pentagon in the xy-plane with each point distance 1 away from the origin, and the
two points on the line perpendicular to the xy-plane through the origin distance 1
away from the origin. To see that any choice of three points gives you an isosceles
triangle, consider the following cases:
Case 1: One of the three points is the origin. Then by construction the other two
points are each distance one away from the origin, and so two sides of the triangle
are congruent. Note that when the other two points are .(0, 0, 1) and .(0, 0, −1),
the triangle is degenerate.
Case 2: All three points are on the pentagon. Then either one point is a vertex
adjacent to the other two vertices on the pentagon, or one point is a vertex
opposite to the other two vertices on the pentagon. In either case, the distance
between that vertex and the other vertices is the same, giving us again an isosceles
triangle.
Case 3: One of the vertices is outside the xy-plane, the two other vertices are
on the pentagon.
√ Then by the Pythagorean theorem, both points on the pentagon
are distance . 2 away from the third vertex, which gives us an isosceles triangle.
Case 4: One vertex is on the pentagon,
√ and the other two are off the xy-plane.
Then the point on the plane is . 2 away from the points off the plane, which
again gives us an isosceles triangle.

S2016-4
(a) Deﬁne .h : A × [0, 1] → A by .h(a, t) = (1 − t)f (a) + tg(a) = (1 − t)a 2 + ta 3 .
Note that .h(a, t) ∈ A = [0, 1] for all .a ∈ A = [0, 1] and all .t ∈ [0, 1].
Moreover, h is a continuous function such that .h(a, 0) = (1−0)f (a)+0g(a) =
f (a) and .h(a, 1) = (1 − 1)f (a) + 1g(a) = g(a), so f and g are homotopic.
(b) • Reflexivity: Suppose .f : A → A is any continuous function on A. Deﬁne
.h : A × [0, 1] → A by .h(a, t) = f (a). Then h is a continuous function such

that .h(a, 0) = f (a) and .h(a, 1) = f (a) for all .a ∈ A, so f is homotopic to

f.
• Symmetry: Suppose .f, g : A → A are any two continuous functions on A
such that f is homotopic to g. This implies that there exists a continuous
154 39 2016 Solutions

function .h : A × [0, 1] → A such that .h(a, 0) = f (a) and .h(a, 1) = g(a)

for all .a ∈ A. Deﬁne .H : A × [0, 1] → A by .H (a, t) = h(a, 1 − t). Then H
is a continuous function such that .H (a, 0) = h(a, 1 − 0) = h(a, 1) = g(a)
and .H (a, 1) = h(a, 1 − 1) = h(a, 0) = f (a) for all .a ∈ A. Therefore, g is
homotopic to f .
• Transitivity: Suppose that .f, g, i : A → A are any three continuous functions
on A such that f is homotopic to g and g is homotopic to i. This implies that
there exist continuous functions .h : A × [0, 1] → A and .h' : A × [0, 1] → A
such that .h(a, 0) = f (a), .h(a, 1) = g(a), .h' (a, 0) = g(a), and .h' (a, 1) =
i(a) for all .a ∈ A. Since .h(a, 1) = g(a) = h' (a, 0) for all .a ∈ A, the
function .H : A × [0, 1] → A deﬁned by
⌠
h(a, 2t) if 0 ≤ t ≤ 12
H (a, t) =
.
'
h (a, 2t − 1) if 12 ≤ t ≤ 1

is continuous, and .H (a, 0) = h(a, 2 · 0) = h(a, 0) = f (a) and .H (a, 1) =

h' (a, 2 · 1 − 1) = h' (a, 1) = i(a) for all .a ∈ A. Therefore, f is homotopic
to i.

S2016-5
(a) .T : R2 → R2 given by .T (x, y) = (y, x + y) is a linear transformation such that
.T (fn , fn+1 ) = (fn+1 , fn + fn+1 ) = (fn+1 , fn+2 ) for all integers .n ≥ 0. We

will prove using induction that .T n (0, 1) = (fn , fn+1 ) for all integers .n ≥ 1. For
the base case, note that .T 1 (0, 1) = T (0, 1) = (1, 0 + 1) = (1, 1) = (f1 , f2 ).
Now assume that .T k (0, 1) = (fk , fk+1 ) for some integer .k ≥ 1, and consider
.T
k+1 (0, 1). Note that

T k+1 (0, 1) = T (T k (0, 1))

= T (fk , fk+1 ) by the inductive hypothesis

= (fk+1 , fk + fk+1 )
= (fk+1 , fk+2 ).

Thus, we can conclude that .T n (0, 1) = (fn , fn+1 ) for all integers .n ≥ 1.
(b) Since .T (1, 0) = (0, 1 + 0) = (0, 1) and since .T (0, 1) = (1, 0 + 1) = (1, 1),
the standard matrix A of T is
⎾ ⏋
01
.A = .
11

The characteristic polynomial

√
of A is .(0 −
√
λ)(1 − λ) − 1 = λ2 − λ − 1, so the
eigenvalues of A are .λ1 = 1+2 5 , λ2 = 1−2 5 . Note that .λ1 = ϕ is the golden
ratio, while .λ2 = 1 − ϕ.
39 2016 Solutions 155

⎾ ⏋ √
1
(c) The vector .v1 = is an eigenvector corresponding to .λ1 = ϕ = 1+ 5
2 since
ϕ

⎾⏋ ⎾ ⏋ ⎾ ⏋
ϕ 1 1
.Av1 = =ϕ = ϕ = ϕv1 ,
1+ϕ 1 + ϕ1 ϕ

⎾ ⏋
1
and similarly, .v2 = is an eigenvector corresponding to .λ2 = 1 − ϕ.
1−ϕ
Note that
⎾ ⏋
0 1
. = √ (v1 − v2 ) .
1 5

Therefore,
⎾ ⏋
0 1 ( )
An
. = √ ϕ n v1 − (1 − ϕ)n v2
1 5
⎛ ⎾ ⏋ ⎾ ⏋⎞
1 1 1
= √ ϕn − (1 − ϕ)n
5 ϕ 1−ϕ
⎾ ⏋
1 ϕ n − (1 − ϕ)n
=√
5 ϕ
n+1 − (1 − ϕ)n+1 .

Since A is the standard matrix of T , this implies that

1 ⎛ ⎞
T n (0, 1) = √ ϕ n − (1 − ϕ)n , ϕ n+1 − (1 − ϕ)n+1 .
.
5

By part (a), the ﬁrst coordinate of .T n (0, 1), namely . √1 (ϕ n − (1 − ϕ)n ), equals
5
.fn . This provides us our desired non-recursive expression for .fn :

1 ( )
fn = √ ϕ n − (1 − ϕ)n .
. (39.1)
5

Remark from the Editors Of course (39.1) is not the only possible non-recursive
expression, and elementary properties of .ϕ such as .1 − ϕ = −1/ϕ lead to
equivalent formulas. If one considers matrix exponentiation (rather than real number
exponentiation) to be a non-recursive expression, then another solution using the
⎾ ⏋n ⎾ ⏋
01 0
above matrix A is the product .fn = [1 0] (or a similar expression in
11 1
terms of the matrix Q from P2003-6 in Chap. 3).
156 39 2016 Solutions

S2016-6
(a) Let .a, b be any two real numbers with .a < b. By the Mean Value Theorem,
there exists a real number c such that .a < c < b and

f (b) − f (a)
f ' (c) =
. .
b−a

Since .f ' (c) = 0, this implies that .0 = f (b)−f

b−a
(a)
, so that .f (a) must equal
.f (b). Since a and b are arbitrary real numbers, f must be a constant function.

(b) To see that this is false, suppose that .A = (0, 1) ∪ (2, 3), and deﬁne .f : A → R
by
⌠
1 if 0 < a < 1,
f (a) =
.
2 if 2 < a < 3.

Then .f ' (a) = 0 for all a in A, but f is not a constant function.

S2016-7
(a) Assume to the contrary that a and b are two elements of such a group G such
that .a /= b but .a 3 = b3 . Note that .(b3 )−1 = (b−1 )3 because

b3 (b−1 )3 = (bb−1 )3 = e3 = e.
.

Therefore, .a 3 = b3 implies that

a 3 (b3 )−1 = b3 (b3 )−1

a 3 (b−1 )3 = e
(ab−1 )3 = e by property (i) of G.

Since there are no elements of G of order 3 (by property (ii) of G) and

(ab−1 )3 = e, it must be the case that .ab−1 = e, which implies that .a = b,
.

which contradicts our assumption that .a /= b. Therefore, it must be the case that
.a /= b whenever .a /= b.
3 3

(b) Assume that G contains n elements and that .φ : G → G is deﬁned by .φ(g) =

g 3 . Note that the previous part implies that .φ is one-to-one since .φ(a) = a 3
equals .φ(b) = b3 if and only if .a = b. We will show that .φ must be onto via
contradiction. Assume that .φ is not onto. Then there exists .g ∈ G such that g is
not in the image of .φ. Then the image of the n elements of G must lie inside the
set .G \ {g} that only contains .(n − 1) elements. Then the pigeonhole principle
implies that there must exist some .g ' ∈ G \ {g} and two distinct elements .a, b
in G such that .φ(a) = g ' = φ(b). However, this is impossible since .φ is one-
39 2016 Solutions 157

to-one. Therefore, it must be the case that .φ is onto. Since .φ is one-to-one and
onto, .φ is a bijection.

S2016-8 Consider, for example, the following sample preference table for 27 voters
in an election with 5 candidates: A, B, C, D, and E:

Rank/number of votes 2 6 7 8 4
First C D B A C
Second E C E D B
Third D E C E E
Fourth A B D C D
Fifth B A A B A

• Plurality—A wins with the most first place votes (eight of them)
• Plurality with Elimination—no candidate has more than half the first place votes,
so we eliminate the candidate(s) with the fewest first place votes. Candidate E is
eliminated. The new preference table is

Rank/number of votes 2 6 7 8 4
First C D B A C
Second D C C D B
Third A B D C D
Fourth B A A B A

Still, no candidate has more than half the ﬁrst place votes, so we eliminate
the candidate(s) with the fewest. Candidates C and D are eliminated. The new
preference table is

Rank/number of votes 2 6 7 8 4
First A B B A B
Second B A A B A

B now has more than half of the ﬁrst place votes, so B is the winner.
• Borda count—the point totals are as follows:
– A: 2(2)+6(1)+7(1)+8(5)+4(1)=61
– B: 2(1)+6(2)+7(5)+8(1)+4(4)=73
– C: 2(5)+6(4)+7(3)+8(2)+4(5)=91
– D: 2(3)+6(5)+7(2)+8(4)+4(2)=90
– E: 2(4)+6(3)+7(4)+8(3)+4(3)=90
Therefore, C wins.
158 39 2016 Solutions

• Pairwise Comparison—the relevant matchups are as follows:

– D vs. A: D wins 19–7 because 19 voters like D better than A.
– D vs. B: D wins 16–11 because 16 voters like D better than B.
– D vs. C: D wins 14–13 because 14 voters like D better than C.
– D vs. E: D wins 14–13 because 14 voters like D better than E.
Therefore, since D went undefeated and untied and since all other candidates lost
at least once (to D), D wins.
• Survivor—we eliminate the candidate(s) with the most last place votes. Candi-
date A is eliminated. The new preference table is

Rank/number of votes 2 6 7 8 4
First C D B D C
Second E C E E B
Third D E C C E
Fourth B B D B D

Now B is eliminated, so the new preference table is

Rank/number of votes 2 6 7 8 4
First C D E D C
Second E C C E E
Third D E D C D

Now D is eliminated, so the new preference table is

Rank/number of votes 2 6 7 8 4
First C C E E C
Second E E C C E

Now C is eliminated, so E wins.

Remark from the Editors Considering the length and complexity of this problem,
one strategy during a timed competition would be to aim for partial credit by
constructing a table comparing only three or four of the voting methods instead
of all ﬁve.
Chapter 40
2017 Solutions

At the .52nd ICMC event at Earlham College, the organizers loaned each team a
straightedge and compass.
The 2017 problem set was prepared by Professor Joshua Cole at St. Joseph’s
College and Professor Stacy Hoehn at Franklin College.
The 2017 problem statements begin on page 45 in Chap. 17.
S2017-1
(a) Draw a circle centered at A that passes through B, as well as a circle centered
at B that passes through A. These two circles will intersect in two points which
we will call C and D. Then draw the line segment connecting C and D. This
line segment will intersect .AB in a point that we will call M. M is the midpoint
of .AB. See the ﬁgure below.

(b) Use the straightedge to extend .AB to the left. Then draw the circle centered at
←→
A that passes through B. This circle will intersect .AB at B and another point.
Draw a circle centered at this new point that passes through B, as well as a circle
centered at B that passes through this new point. Then draw the line segment
that connects the two points where these circles intersect. This line segment
passes through A and is perpendicular to .AB. The corner D of our square will

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 159
A. Coffman et al., The Indiana College Mathematics Competition (2001–2023),
Problem Books in Mathematics, https://doi.org/10.1007/978-3-031-62768-2_40
160 40 2017 Solutions

be one of the points where this line segment intersects the circle centered at A
that passes through B. Finally, draw a circle centered at D that passes through A
and a circle centered at B that passes through A. These two circles will intersect
at A and another point, namely the ﬁnal corner C of our square. To complete
the square, draw the line segments connecting C with D and C with B. See the
ﬁgure.

If T , R, and L denote the areas of the indicated regions in the above ﬁgure,
T = 12 s 2 since it is the area of a triangle with base s and height s,
.

⎛ √ ⎞2
1 2s π
.T + 2R = π = s2
2 2 4
√
since .T + 2R is the area inside a semicircle with radius . 2s
2 , and

1 ⎛ s ⎞2 π
.L+R = π = s2
2 2 8
40 2017 Solutions 161

since .L+R is the area inside a semicircle with radius . 2s . We can use the ﬁrst two
equations to see that .R = π8 s 2 − 41 s 2 . We can then use this value for R as well
as the third equation given above to see that L, which is the area of the lune, is
.L = s . Thus, the area of the lune will equal the area of a square that has side
1 2
4
length . 12 s, which is just the length of the line segment from B to the midpoint
of .AB. Therefore, to construct a square with area equal to the area of one of the
lunes, we can use our construction from part (a) to construct the midpoint M of
.AB and then use our construction from part (b) to construct a square that has

.MB as one of its sides. (Alternatively, since triangle ABC is a right isosceles

triangle, the polygon that connects B, the midpoint M of .AB, the midpoint N
of .AC, and the midpoint O of .BC will be a square with side length . 12 s.)

S2017-2 This series does converge. To see this, we make a comparison to the series
⎲ ∞ ⎛ ⎞n
1
. , which is a convergent geometric series, since its ratio is . 21 < 1. We
2
n=1 ⎛ ⎞n
demonstrate that .0 ≤ (an )n ≤ 12 for all n. If n is odd, then .an is the .nth partial
⎲∞ ⎛ ⎞n
1 1 1
sum of the geometric series . , which converges to . 3 1 = 32 = 21 . Since the
3 1− 3 3
n=1
terms of that series are nonnegative, it follows that .0 ≤ an ≤ 12 . Thus .0 ≤ (an )n ≤
⎛ ⎞n
2 , as we desire for the comparison test. If n is even, then .an = | sin(n) cos(n)| =
1
⎛ ⎞n
1
2 | sin(2n)|. Since .| sin(2n)| ≤ 1 for all n, .0 ≤ an ≤
1
2 . Then .0 ≤ (an ) ≤
n 1
2
for all n, as we desire for the comparison test. We have shown that for all n (odd or
⎛ ⎞n ⎲∞
even), .0 ≤ (an ) ≤ 2 . Hence, by the comparison test, . (an )n converges.
n 1

n=1

S2017-3 Note that .x

2+ ax + b = 0 has at least one real root as long as the
discriminant, which is .a 2 − 4b in this case, is nonnegative. Thus, we need to find the
probability that .a 2 − 4b ≥ 0, i.e., .b ≤ 41 a 2 . Since each student is picking a number
from .[0, 1] according to the uniform probability distribution, finding the desired
probability is equivalent to finding the proportion of .[0, 1] × [0, 1], where .b ≤ 41 a 2 .
162 40 2017 Solutions

This region is shaded in the ﬁgure below, where the horizontal axis corresponds to
a and the vertical axis corresponds to b.

The area of the shaded region is

⎰ 1 1 2 1
. a da = ,
0 4 12

while the area of the entire square is 1. Thus, the desired probability is the ratio of
1
these areas, namely . 12 .
S2017-4
(a) Suppose that the diagram below represents a .3 × 3 multiplicative magic square
with magic product P .

a b c
d e f
g h i

Then .P = abc = def = ghi = adg = beh = cf i = aei = ceg.

P 4 = (aei)(ceg)(def )(beh) = (abc)(def )(ghi)e3 = P 3 e3 ,

and the conclusion is .P = e3 .

Remark from the Editors Part (a) repeats P2003-4; see S2003-4 in Chap. 26 for an
alternate solution.

(b) By the previous part, we know that the magic product must be a perfect cube.
Moreover, the magic product must have at least nine distinct divisors because
each entry in the magic square is a divisor of the magic product. The ﬁrst ﬁve
perfect cubes, 1, 8, 27, 64, and 125, have 1, 4, 4, 7, and 4 divisors, respectively.
The next perfect cube, 216, has 16 divisors, so it is a candidate for the minimal
40 2017 Solutions 163

magic product. Since the sample array given below is a .3 × 3 multiplicative

magic square with magic product 216, 216 is, in fact, the minimal magic
product.

12 1 18
9 6 4
2 36 3

Other multiplicative magic squares with magic product 216 are possible.
However, our solution to the previous part implies that all of them will have
a 6 in the middle square.
S2017-5
(a) This formula relies on the fact that for any positive integer k, the sum of the
integers from 1 to k is . 12 k(k + 1). We will ﬁrst consider the case when n is odd.
In this situation,

n−1 n−1
s(n) = 0 + 1 + 1 + 2 + 2 + · · · +
. +
2 2
⎛ ⎞
n−1
= 2 1 + 2 + ··· +
2
⎛ ⎛ ⎞⎛ ⎞⎞
1 n−1 n−1
=2 +1
2 2 2
⎛ ⎞⎛ ⎞
n−1 n+1
=
2 2
n2 − 1
= .
4
When n is even,
⎛n ⎞ ⎛n ⎞ n
s(n) = 0 + 1 + 1 + 2 + 2 + · · · +
. −1 + −1 +
2 2 2
⎛ ⎛n ⎞⎞ ⎛ ⎛n ⎞ n⎞
= 1 + 2 + ··· + − 1 + 1 + 2 + ··· + −1 +
2 2 2
⎛ ⎛ ⎞ ⎛ ⎞ ⎞ ⎛ ⎛ ⎞⎛ ⎞⎞
1 n n 1 n n
= −1 + +1
2 2 2 2 2 2
⎛ ⎛ ⎞ ⎞ ⎛ ⎛ ⎞⎛ ⎞⎞
1 n − 2 ⎛n⎞ 1 n n+2
= +
2 2 2 2 2 2
⎛ 2 ⎞ ⎛ 2 ⎞
n − 2n n + 2n
= +
8 8
164 40 2017 Solutions

n2
= .
4
Thus,
⌠
n2 −1
if n is odd,
s(n) =
.
n2
4
4 if n is even.

(b) Note that .m + n and .m − n differ by 2n, which is even. Thus, .m + n and .m − n
are either both odd or both even. In the case when they are both odd,

(m + n)2 − 1 (m − n)2 − 1
s(m + n) − s(m − n) =
. −
4 4
(m2 + 2mn + n2 − 1) − (m2 − 2mn + n2 − 1)
=
4
4mn
=
4
= mn.

Meanwhile, in the case when they are both even,

(m + n)2 (m − n)2
s(m + n) − s(m − n) =
. −
4 4
(m2 + 2mn + n2 ) − (m2 − 2mn + n2 )
=
4
4mn
=
4
= mn.

S2017-6 We will give a proof by induction. For the base case, we will consider the
case when .n = 1. Then .A = (1) or .A = (−1), so .det(A) equals 1 or .−1, both of
which are divisible by .2n−1 = 20 = 1. For the inductive hypothesis, assume that for
some .k ≥ 1, the determinant of any .k ×k matrix whose entries are all .±1 is divisible
by .2k−1 . For the inductive step, suppose that A is any .(k + 1) × (k + 1) matrix such
that every entry of A is .±1. Let B be the .(k + 1) × (k + 1) matrix obtained from
A by replacing row 1 of A by the sum of row 1 with row 2. Since each entry of A
is 1 or .−1, the entries in the first row of B must all be 2, 0, or .−2, while the entries
in the remaining rows of B must all be 1 or .−1. Moreover, .det(A) = det(B) since
the type of elementary row operation performed does not change the determinant.
Therefore, it suffices to show that .det(B) is divisible by .2(k+1)−1 = 2k . By using
cofactor expansion along the first row of B, we see that
40 2017 Solutions 165

⎲
k+1
. det(B) = (−1)1+j b1j det(M1j ),
j =1

where .M1j is the .k × k matrix obtained from B by removing row 1 and column
j . For each j , .M1j is a .k × k matrix whose entries are all 1 or .−1, so by the
inductive hypothesis, .det(M1j ) is divisible by .2k−1 for all j . Moreover, for each j ,
.b1j is divisible by 2 since .b1j equals 2, 0, or .−2. Thus, for each j , .b1j det(M1j ) is

divisible by .2 · 2k−1 = 2k , so

⎲
k+1
. det(B) = (−1)1+j b1j det(M1j )
j =1

is also divisible by .2k . Since .det(A) = det(B), .det(A) is divisible by .2k , completing
our induction.
S2017-7 Let .σ denote the permutation implemented by the robot. Then in cycle
notation,

. σ 7 = (A 7 J 9 10 5 Q 6 3 K 8 4 2).

Thus, .|σ 7 | = 13. We claim that this implies that .|σ | = 13 as well. To see this, note
that .|σ 7 | = 13 implies that .(σ 7 )13 = e, where e denotes the identity permutation.
Then .σ 91 = e. Thus it implies that .|σ | divides 91, meaning that .|σ | is 1, 7, 13, or 91.
But .|σ | cannot equal 1 or 7 because then .σ 7 would have to equal e, which it does
not. Moreover, if .|σ | was 91, then the least common multiple of the lengths of the
cycles when .σ is written in disjoint cycle notation would have to be 91, which would
mean that .σ would have disjoint cycles of lengths 7 and 13 or a cycle of length 91,
both of which are impossible with only a total of 13 cards to permute. Therefore, .σ
must have order 13. Since .σ 13 = e, .(σ 7 )2 = σ 14 = σ σ 13 = σ e = σ . Hence, in
cycle notation,

σ = (σ 7 )2 = (A J 10 Q 3 8 2 7 9 5 6 K 4),
.

so the order of the cards after the ﬁrst shufﬂe is

. 4, 8, Q, K, 9, 5, 2, 3, 7, J, A, 10, 6.

S2017-8 First note that A cannot be finite. For a finite set A, around each element
a of A, there is an open interval small enough so that it contains a but no other
element of A, contradicting the given condition on A. Thus, A must be infinite.
Next, we will show that A cannot be countably infinite via contradiction.
Suppose that A is countably infinite. Then the elements of A can be listed out as
.a1 , a2 , a3 , . . . . Let .U1 be the open interval .(a1 − 1, a1 + 1) around .a1 . By the
166 40 2017 Solutions

condition on A, this open interval must contain another element in A; without loss
of generality, assume that .a2 is in .U1 (if not, relabel the elements of A). Next, let .U2
be a small enough open interval around .a2 so that the closure of .U2 (i.e., the closed
interval consisting of .U2 and the corresponding endpoints) is contained inside .U1
and so that .x1 is not in the closure of .U2 . Again, by the condition on A, .U2 must
contain another element of A; without loss of generality, assume that .a3 is in .U2 .
Now choose an open interval .U3 around .a3 such that the closure of .U3 is contained
in .U2 and so that neither .a1 nor .a2 is contained in the closure of .U3 . Without loss of
generality, assume that .a4 is in .U3 . In general, for each .n ≥ 2, ﬁnd a small enough
open interval .Un around .an such that:
• The closure of .Un is contained in .Un−1 .
• .a1 , a2 , . . . , an−1 are not contained in the closure of .Un .
• .an+1 is contained in .Un (after a possible relabeling of the points in
.{an+1 , an+2 , . . . }).

For each .n ≥ 1, let .Vn = closure(Un ) ∩ A. Note that each .Vn is closed and bounded,
∞
⋂
hence compact. Moreover, .V1 ⊃ V2 ⊃ V3 ⊃ · · · . Let .V = Vn . Since V is the
i=1
intersection of a decreasing nested sequence of non-empty compact subsets of A, V
is a non-empty subset of A. However, none of the .an ’s can be in V since for .i ≥ n,
.an is not in .Vi . Since we have arrived at a contradiction, A cannot be countably

inﬁnite. Thus, it must be the case that A is uncountably inﬁnite.

Chapter 41
2018 Solutions

The .53rd contest, temporarily renamed the Intersectional Collegiate Mathematics

Competition, was held at Valparaiso University, during the Tri-Section meeting of
the Illinois, Indiana, and Michigan Sections of the Mathematical Association of
America. Over 50 teams, from all three states, participated.
The 2018 problem set was prepared by Professors Daniel Maxin, Zsuzsanna
Szaniszlo, and Tiffany Kolba at the host institution.
The 2018 problem statements begin on page 49 in Chap. 18.

S2018-1 For the claimed equality to happen, one needs .| sin(j x)| = 1 for every
j = 1, . . . , n. Since .n ≥ 2, we have
.

| sin(x)| = 1 and | sin(2x)| = 1,

which leads to .x = π2 + kπ and .2x = π2 + 𝓁π with k, .𝓁 integers. Combining the

two, one obtains .𝓁 − 2k = 12 , which is a contradiction.

S2018-2 Rationalizing the denominators, one obtains the following telescopic

summation:
√ √ √ √ √
( 2 − 1) + ( 3 − 2) + · · · + ( n + 1 − n) ≥ 100,
.

which simplifies to
√
. n + 1 − 1 ≥ 100.

Hence the smallest n is 10200.

S2018-3 Consider the function

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 167
A. Coffman et al., The Indiana College Mathematics Competition (2001–2023),
Problem Books in Mathematics, https://doi.org/10.1007/978-3-031-62768-2_41
168 41 2018 Solutions

f (x) = a1x + a2x + · · · + anx .

Notice that .f (0) = n and, from the hypothesis, .f (x) ≥ f (0) = n. Therefore .f (x)
has a local minimum at .x = 0. From Fermat’s theorem, .f ' (0) = 0. But,

f ' (x) = a1x ln(a1 ) + · · · + anx ln(an ).

Then

f ' (0) = ln(a1 ) + · · · + ln(an ) = ln(a1 a2 · · · an ) = 0

and .a1 a2 · · · an = 1.

S2018-4 First, we use the following row and column operations, which will not
change the value of the determinant: Add rows 2 and 3 to row 1. Then, subtract
column 1 from column 2 and column 1 from column 3. With these operations, the
determinant becomes
⎡ ⎤
x+y+z 0 0
. det ⎣ y x −y z−y⎦
z y−z x−z
⎡ ⎤
1 0 0
= (x + y + z) det ⎣ y x − y z − y ⎦
z y−z x−z
= (x + y + z)(x 2 + y 2 + z2 − xy − yz − zx)
1
= (x + y + z)((x − y)2 + (y − z)2 + (z − x)2 ) ≥ 0.
2

S2018-5 Notice
1 1 1 1 1 1
. = − , or = − .
xk+1 xk xk + 1 xk + 1 xk xk+1

Adding up for .k = 1, . . . , 100, we get a telescoping cancelation and

1 1 1
S100 =
. − =2− .
x1 x101 x101

Since .xn is increasing and .x3 > 1, we have .0 < 1

x101 < 1 and .S100 = 2 − 1
x101 < 2.
So .⎿S100 ⏌ = 1.

S2018-6A First notice that, by dividing by .|AB|2 , the equality to prove becomes
41 2018 Solutions 169

|AM| |AD| |BM| |BE|

. + = 1.
|AB| |AB| |AB| |AB|

Consider the perpendicular from M to AB and denote by P the intersection of this

perpendicular with AB.
Notice that the triangles AMP and ABD are similar and the ratio of the segments
opposite the congruent angles is the same. Hence

|AM| |AP |
. = .
|AB| |AD|

Analogously, notice that the triangles BMP and BAE are also similar and, from
there, we infer that

|BM| |P B|
. = .
|AB| |BE|

Substituting these into the equality to prove, we obtain

|AP | |AD| |P B| |BE| |AP | + |P B| |AB|

. + = = = 1.
|AD| |AB| |BE| |AB| |AB| |AB|

S2018-6B Using the fact that the inscribed triangles .ΔABD and .ΔABE have right
angles at D and E,

. |AM||AD| + |BM||BE|
= |AM|(|AM| + |BM| cos / BMD) + |BM|(|BM| + |AM| cos / AME)
= |AM|2 + |BM|2 + 2|AM||BM| cos / AME
= |AM|2 + |BM|2 − 2|AM||BM| cos / AMB
= |AB|2 .

The last step uses the Law of Cosines.

S2018-7 Let .Y = the number of coin flips. On average, it takes 2 flips to observe a
head. Hence, on average, it takes 2X flips to observe X heads. Then, we have

E[Y ] = E[E[Y |X]] = E[2X] = 2E[X] = 2(3.5) = 7.

For a more direct computation, note that all outcomes of the die roll have probability
1
6 . For each of those equally weighted cases, one would expect to flip a coin 2, 4, 6,
.

8, 10, and 12 times, respectively. Thus,

170 41 2018 Solutions

. E[Y ]
1 1 1 1 1 1
= × 2 + × 4 + × 6 + × 8 + × 10 + × 12
6 6 6 6 6 6
1
= (2 + 4 + 6 + 8 + 10 + 12)
6
= 7.

S2018-8 Let .c ≥ 0 be the number of clubs. Since each pair of clubs (but not three)
shares a member, the number
⎛ ⎞ of pairs of clubs cannot be more than the number of
c
people. In other words, . ≤ n. This condition becomes
2

c2 − c − 2n ≤ 0,
.

which is a quadratic in c, concave upward with two real roots, one positive and one
nonpositive. Therefore c must be bounded by the positive root, i.e.,
√
0 ≤ c ≤ (1 +
. 1 + 8n)/2. (41.1)

Claim: The number of clubs that may be formed

√ can be any integer satisfying (41.1).
In particular, the maximum number, .[(1 + 1 + 8n)/2] (where .[x] denotes the
integer part of .x > 0), can be achieved.
Note that for any number (.n ≥ 0) of townspeople, a town with no clubs (.c = 0)
or one club (.c = 1) meets, vacuously, the criteria of the problem, and it is also
permitted that there are exactly two clubs (.c = 2) as long as there is at least one
person who is a member of both (.n ≥ 1).
To establish the claim for any .c ≥ 3 satisfying (41.1), we assume there are c clubs
and need to show that there exists an assignment of members to clubs satisfying the
criteria. Because .c(c − 1)/2 ≤ n, there exists a one-to-one function f from the list
of .c(c − 1)/2 pairs of clubs to the list of .n ≥ 3 people. For a pair of clubs, .{cj , ck },
.j /= k, if .f ({cj , ck }) = A, then assign person A to be a member of .cj and of .ck ,

and no other clubs. If person B is not in the image of f , then B belongs to no clubs.
Now each club .ck has at least one member .f ({cj , ck }) in common with any other
club .cj , and by construction there is no person who is a member of more than two
clubs.
For example, if .n = 6, then the maximum is .c = 4. Let the people be denoted by A,
B, C, D, E, F , the clubs by 1, 2, 3, 4. Based on the assignment:

A in 1, 2 B in 1, 3 C in 1, 4 D in 2, 3 E in 2, 4 F in 3, 4,
.

the members of the clubs are as follows:

41 2018 Solutions 171

1 : A, B, C 2 : A, C, E 3 : B, C, F 4 : D, E, F.
.

√
For another example, if .n = 5, then the maximum is .c = [(1+ 41)/2] = 3. Let the
people be denoted by A, B, C, D, E, the clubs by 1, 2, 3. Based on the assignment:

A in 1, 2
. B in 1, 3 C in 2, 3,

the members of the clubs are as follows:

. 1 : A, B 2 : A, C 3 : B, C,

and D and E do not belong to any club. This meets the criteria. If there were a fourth
club, then the number of pairs of clubs would be 6 as in the previous example, so
by the first criterion, one of the five people must belong to two distinct pairs of
clubs—either three or four clubs, a violation of the second criterion.
Chapter 42
2019 Solutions

The .54th competition was held at the University of Indianapolis.

The 2019 problem set was prepared by Professor Paul Fonstad at Franklin College.
The 2019 problem statements begin on page 51 in Chap. 19.

S2019-1A

(a) Let .f (x) = x + sin(x) and .g(x) = −x. Then f and g are both continuous.
Individually, each function is monotone, but the sum .(f + g)(x) = sin(x) is
not.
(b) The function

sin( x1 )
f (x) =
.
x
is not bounded in any neighborhood of .x = 0, and neither of the one-sided
limits at .x = 0 exists.
(c) Consider the function
⌠
0 if x is rational,
f (x) =
.
x if x is irrational.

This .f (x) is continuous at .x = 0 but discontinuous everywhere else.

(d) The function .f (x) = x 3 has .f ' (0) = 0, yet the function is increasing on .R.
(e) If .f (x) = 1/x on the domain .(−∞, 0) ∪ (0, ∞), then .F (x) = ln |x| is an
antiderivative of f , satisfying .F ' (x) = f (x) at every point in the domain. F is
⎰ 1
1
defined at both .x = −1 and .x = 1, but . dx does not converge.
−1 x

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 173
A. Coffman et al., The Indiana College Mathematics Competition (2001–2023),
Problem Books in Mathematics, https://doi.org/10.1007/978-3-031-62768-2_42
174 42 2019 Solutions

S2019-1B For part (e), there is another counterexample with .f (x) defined on all of
⎰ 1
'
.R and .F (x) satisfying .F (x) = f (x) at every point, but where . f (x)dx fails to
−1
exist as a proper Riemann integral because f is unbounded near .x = 0.
⌠
|x|3/2 sin(1/x) for x /= 0,
.F (x) =
0 for x = 0.
⎧ 3 1/2
⎨ 2 |x| sin(1/x) − |x|−1/2 cos(1/x) for x > 0,
f (x) = 0 for x = 0,
⎩ 3 1/2 −1/2
− 2 |x| sin(1/x) − |x| cos(1/x) for x < 0.
⎰ 1
The integral . f (x)dx could be evaluated using some other method (an improper
−1
integral, or some method of integration other than Riemann’s) to give .F (1)−F (−1).

S2019-2 All such triangles must pass through the point .(0, 4). To see this, first note
that since the right angle occurs at the origin, the legs of the right triangle are formed
by perpendicular lines whose equations can be written as .y = mx and .y = −x/m
for some .m > 0. These lines intersect the parabola at the points .(4m, 4m2 ) and
2
.(−4/m, 4/m ), respectively. Thus, the equation for the hypotenuse of the triangle

is
⎛ 2 ⎞
4m − m42
.y = (x − 4m) + 4m2
4m + m4
⎛ 2 ⎞
4m − m42
= x + 4.
4m + m4

No matter what m is, when x equals 0, y will equal 4. Thus, the hypotenuse of the
right triangle will always pass through the point .(0, 4).

S2019-3

(a) Note that the amount of money that party guest n makes (or loses) is given by
.f (n) = 10(en − on ), where .en denotes the number of even positive divisors of

n and .on denotes the number of odd positive divisors of n (including 1). Thus, if
n is odd, guest n will lose money since .en = 0. If n is even but not divisible by
4, guest n will break even, since every even divisor of n can be paired with an
odd divisor of n. Finally, if n is divisible by 4, party guest n will make money
since n will have more even divisors than odd divisors since every odd divisor
is paired with an even divisor but not every even divisor is paired with an odd
divisor. Thus, 30 guests will lose money, 15 will break even, and 15 will make
money.
42 2019 Solutions 175

(b) By the previous part, we only need to consider the 15 guests whose number n is
divisible by 4. The party guest who makes the most money will correspond to
the number n for which the number of even divisors exceeds the number of odd
divisors by the most. If n is a power of 2, .n = 2k , then .f (n) = 10(k − 1), for a
maximum of .f (32) = 40. When n has a prime factorization of the form .22 p, for
p an odd prime (3, 5, 7, 11, 13), .f (n) = 10(4 − 2) = 20. Similarly, when n has
a prime factorization of the form .23 p (for .p = 3, 5, 7), .f (n) = 10(6 − 2) = 40.
The only numbers n remaining are .f (36) = f (22 32 ) = 10(6 − 3) = 30,
.f (48) = f (2 3 ) = 10(8 − 2) = 60, and .f (60) = f (2 3 5 ) = 10(8 − 4) =
4 1 2 1 1

40. Thus, the .48th party guest in line will make the most money.
(c) By part (a), we only need to consider guests whose number n is odd. The party
guest who loses the most money will correspond to the number n which has
the most odd divisors. .f (1) = 10(0 − 1) = −10. Considering odd primes
.1 < p < q, .f (p) = 10(0 − 2) = −20, .f (p ) = 10(0 − 3) = −30, and
2

.f (pq) = 10(0 − 4) = −40. The only numbers n remaining are .f (27) =

f (33 ) = 10(0 − 4) = −40 and .f (45) = f (32 51 ) = 10(0 − 6) = −60. So, the
guest who lost the most would be the .45th guest in line.
(d) The sum of positive f values from part (b) is

(10 + 20 + 30 + 40) + 20 · 5 + 40 · 3 + 30 + 60 + 40 = 450.

The sum of negative f values from part (c) is

. − 10 − 20 · 16 − 30 · 3 − 40 · 8 − 40 − 60 = −840.

⎲60
Since . f (n) = −390, Two-Face and his henchmen made .$390. Note that by
n=1
part (a), you only need to consider those values of n that are odd or that are
divisible by 4 when computing this sum.

Remark from the Editors This problem and solution are similar to ICMC P1996-3
[AFMC].

S2019-4

(a) Let X denote the number of flips needed to get your first tail. Flipping a H on
your first toss adds one to your required number of tosses and gets you no closer
to your goal. Meanwhile, flipping a T on your first toss completes your task in
a single toss. Thus,

E(X) = P (H )E(X|H ) + P (T )E(X|T )

= P (H )(E(X) + 1) + P (T )(1)
176 42 2019 Solutions

1 1
= (E(X) + 1) + .
2 2

This implies . 12 E(X) = 1, which implies .E(X) = 2.

Alternatively,
∞
⎲ ⎛ ⎞n
1
.E(X) = n = 2.
2
n=1

(b) Let Y denote the number of flips needed to achieve your first string of heads
followed by tails. Flipping a T on your first toss adds one to your required
number of tosses and gets you no closer to your goal. If you flip a H on your
first toss, you are just waiting on your first T . Let X denote the number of
additional tosses required to get your first T if your first toss is H . From the
previous part, we know that .E(X) = 2. Then

E(Y ) = P (T )E(Y |T ) + P (H )(E(X) + 1)

1 1
= (E(Y ) + 1) + (2 + 1)
2 2
1
= (E(Y )) + 2.
2

This implies . 12 E(Y ) = 2, which implies .E(Y ) = 4.

Alternatively,
∞
⎲ ⎛ ⎞
n−1
.E(Y ) = n = 4.
2n
n=1

(c) Let Z denote the number of flips needed to achieve the first string of heads
followed by heads. Flipping a T on your first toss adds one to your number
of required flips but does not get you closer to observing two heads. Likewise,
starting out by flipping H T adds two flips to your total but does not get you
closer to observing two heads. Meanwhile, starting out with H H takes two
flips and completes the task in two tosses. Thus,

E(Z) = P (T )E(Z|T ) + P (H T )E(Z|H T ) + P (H H )E(Z|H H )

= P (T )(E(Z) + 1) + P (H T )(E(Z) + 2) + P (H H )(2)

1 1 1
= (E(Z) + 1) + (E(Z) + 2) + (2).
2 4 4

This implies . 14 E(Z) = 32 , which implies .E(Z) = 6.

42 2019 Solutions 177

⎾ ⏋
1 2019n
S2019-5 Claim: For any integer n, .An = .
0 1

⎾ first observe⏋that .A =
We 1
⎾ will prove⏋ this using induction,
⎾ twice.
⏋ For the base cases,
1 2019 · 1 1 2019 · 0 1 2019 · (−1)
, .A0 = I = , and that .A−1 = since
0 1 0 1 0 1
⎾ ⏋⎾ ⏋ ⎾ ⏋
1 2019 1 −2019 10
. = .
0 1 0 1 01

⎾ ⏋
1 2019k
Next, assume that for some positive integer k, .Ak = . Then
0 1
⎾ ⏋ ⎾ ⏋
1 2019k 1 2019
.A
k+1
=A A= k
·
0 1 0 1
⎾ ⏋
1 · 1 + 2019k · 0 1 · 2019 + 2019k · 1
=
0·1+1·0 0 · 2019 + 1 · 1
⎾ ⏋
1 2019(k + 1)
= .
0 1
⎾ ⏋
1 2019k
Finally, assume that for some negative integer k, .Ak = . Then, similarly,
0 1
⎾ ⏋⎾ ⏋
1 2019k 1 −2019
.Ak−1 = Ak A−1 =
0 1 0 1
⎾ ⏋
1 2019(k − 1)
= .
0 1

This completes our proof.

S2019-6 The sequences .(sn )∞ ∞

n=1 and .(tn )n=1 must converge, while not enough
information is provided to determine whether the sequence .(un )∞
n=1 converges.
To see that .(sn )∞
n=1 must converge, note that it is a monotone decreasing sequence
that is bounded below by .t1 , since for all .n ∈ N, .sn ≥ tn ≥ t1 by hypothesis. Every
monotone decreasing sequence that is bounded below converges.
To see that .(tn )∞
n=1 must converge, note that it is a monotone increasing sequence
that is bounded above by .s1 , since for all .n ∈ N, .tn ≤ sn ≤ s1 by hypothesis. Every
monotone increasing sequence that is bounded above converges.
Not enough information is given to determine whether the sequence .(un )∞
n=1 will
converge. There are some examples that satisfy all the given criteria for which
178 42 2019 Solutions

.(un )∞
n=1 will converge, and others for which it will not converge. For example, if
.sn = 1 +
∞
n , .tn = −1 − n , and .un = 0, then .(sn )n=1 is a monotone decreasing
1 1
∞
sequence, .(tn )n=1 is a monotone increasing sequence, and .sn ≥ un ≥ tn for all
.n ∈ N, and .(un )
∞ does converge. Meanwhile, for the same .(s )∞ and .(t )∞ ,
n=1 n n=1 n n=1
the sequence defined by .un = (−1)n also satisfies the given conditions, but .(un )∞
n=1
does not converge.

S2019-7 Since x and y generate G, every element of G can be written as a finite

product of x’s, y’s, .x −1 ’s, and .y −1 ’s.

Step 1. Based on the relation .xx = e, we have .x −1 = x, and based on the relation
.yyyy = e, we have .y
−1 = y 3 , so every element of G can be written as a finite

product of just x’s and y’s, where you never need more than one x in a row or
more than three y’s in a row.
Step 2. We next show that any expression (already simplified as in Step 1.) with three
y’s in a row followed or preceded by an x can be rewritten using the .xyxyxy = e
relation to have a smaller number of y’s, without increasing its total number of
letters:

. y 3 x = y −1 x −1 = (xy)−1 = xyxy.

Similarly, using .xyxyxy = e =⇒ yxyxy = x =⇒ yxyxyx = e,

.xy 3 = x −1 y −1 = (yx)−1 = yxyx.

Then, any expression with three blocks of y’s separated by two x’s (of the form
· · · yxy k xy · · · ) can also be simplified to have a smaller number of y’s and a
.

smaller number of letters:

yxyxy = x,
.

yxy 2 xy = (yxyx)(xyxy) = (xy 3 )(y 3 x) = xy 2 x,

yxy 3 xy = y(yxyx)xy = y 2 xy 2 .

Any product in G has a minimum length expression in x, y as in Step 1, and

among products of the same length, we can consider words with the minimum
number of y’s. By the above calculations, such a minimal word does not have
three consecutive y’s unless it is .y 3 , and does not have three blocks of y’s
separated by two x’s. This leaves the following list (possibly with repeats) of
possible minimal expressions: There are four expressions of the form .y {0,1,2,3} ,
three of the form .xy {0,1,2} , twelve of the form .x {0,1} y {1,2} xy {0,1,2} , and eight of
the form .x {0,1} y {1,2} xy {1,2} x. The conclusion so far is that G has at most 27
elements.
42 2019 Solutions 179

Step 3. If .y = e, then .G = {e, x} and G has at most two elements. If .y 2 = e, then

the list from Step 2. becomes: .G = {e, y, xy {0,1} , x {0,1} yxy {0,1} , x {0,1} yxyx},
with at most 10 elements. If .y 3 = e = y 4 , then .y = e. The remaining case is
that G has a subgroup with exactly four elements, .{0, y, y 2 .y 3 }, and the number
of elements of G is a multiple of 4. In any case, by the upper bound from Step
2., G has at most 24 elements.
Step 4. An example showing that 24 is the maximum that can be achieved is the
symmetric group on four elements, generated by .x = (12) and .y = (1234).

One way to visualize groups is via a Cayley graph that illustrates the relations
between the generators x and y. A Cayley graph for this 24-element group is the
following truncated octahedron:

S2019-8 The faces of the pyramid T are determined by the xy-plane, the xz-plane,
the yz-plane, and the plane P defined by .z = −2x − 3y + 24. Since one corner of
the rectangular prism R must be at the origin, the volume of R will be maximized
when its faces meeting at the origin are also aligned with the coordinate planes,
and the opposite corner is on the plane P . If this corner is located at the point
.(x, y, −2x − 3y + 24) on the plane P , then the volume of R is given by

V (x, y) = xy(−2x − 3y + 24) = −2x 2 2y − 3xy 2 + 24xy.

Taking partial derivatives, we get that

Vx (x, y) = −4xy − 3y 2 + 24y = y(−4x − 3y + 24)

and that

Vy (x, y) = −2x 2 − 6xy + 24x = x(−2x − 6y + 24).

.
180 42 2019 Solutions

Setting these partial derivatives equal to zero and solving for x and y, we see that
(x, y) = (4, 83 ) is one critical point of V . Using the second partial derivative test,
.

we can see that this critical point corresponds to a maximum for V . This means that
the maximum volume would be
8 8 8 1
V = 4( )(−2(4) − 3( ) + 24) = 4( )(8) = 85
.
3 3 3 3
cubic units.
Chapter 43
2021 Solutions

By the spring of 2021 during the pandemic, faculty and students had begun to adapt
to working in virtual and remote environments. The Spring 2021 Indiana MAA
Section meeting was conducted virtually as a one-day Zoom meeting, and the .55th
ICMC happened on an earlier day, organized into a decentralized, remote format,
where each team would meet on its own campus, with a local faculty proctor. After
the teams completed the exam in the allotted amount of time, the proctor scanned
and emailed their solutions to the Section’s Student Activities Coordinator who
distributed them to graders across the state. The results were announced during the
Zoom meeting.
The 2021 problem set was prepared by Professor Harold Reiter at the University of
North Carolina at Charlotte.
The 2021 problem statements begin on page 57 in Chap. 21.

Applause via Zoom for the 2021 competition’s ﬁrst place team from Purdue University West
Lafayette

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 181
A. Coffman et al., The Indiana College Mathematics Competition (2001–2023),
Problem Books in Mathematics, https://doi.org/10.1007/978-3-031-62768-2_43
182 43 2021 Solutions

S2021-1
⎾ ⏋
1 2021 ∗ n
(a) Notice that since A is upper triangular, .A=
n . Hence .A2021 =
0 1
⎾ ⏋
1 20212
.
0 1
(b) We can reverse the previous problem
⎾ ⏋ to see that B can also be chosen to be
11
upper triangular and that .B = works.
01

S2021-2 The polynomial .q(x) must factor into the form .(cx+d)2 (ex+f )2 because
it has a double root at each point where .mx + b is tangent to the graph of .p(x).
Suppose that r and s are these double roots. We use results about the roots, due to
Viète. The sum of the roots, .2r +2s, must equal .−(32/16) = 2. Thus .s = 1−r. The
sums of the products of the roots, .r ∗r +r ∗s +s ∗r +s ∗s = (1−r)2 +r 2 +4r(1−r),
must equal .−104/16. Solving this for r yields .r = −3/2, so .s = 5/2. We can then
ﬁnd m and b. The product of the roots equals .(232 − b)/16, which yields .b = 7. For
m, we can simply evaluate .p' (r) or .p' (s), yielding .m = 2.

S2021-3

(a) Let .P (n) be the product of the digits. If any digit is even, then .P (n) is even.
Since n is also a multiple of 5, it must be a multiple of 10. But then .P (n) = 0,
a contradiction.
(b) The condition above that n must be a multiple of 5, but not a multiple of 10,
means that the one’s place digit must be 5. If a is the hundred’s place digit and
b is the ten’s place digit, then .n = 100a + 10b + 5 = 5 ∗ 5ab. Dividing by
5, .20a + 2b + 1 = 5ab. Rearranging, .5ab − 20a − 2b − 1 = 0. Factoring,
.(5a − 2)(b − 4) = 9. We can then see that .a = 1 and .b = 7 work, yielding

.n = 175.

S2021-4 Let .AB = A − B, etc. Then .||AB||2 = 72, .||AC||2 = 24, and .||BC||2 =
48. This shows that the given vertices are deﬁnitely not all on the same √ face of the
supposed cube. If x is the√edge of a cube, then the diagonal of a face is .x 2 and the
diagonal of the cube is .x 3. The squared norms above are in a .3 : 2 : 1 ratio, which
shows that AC is the edge of the cube, BC is the diagonal of a face, and AB is the
diagonal of the cube.

S2021-5

(a) Start with 1, and pair it with any of the seven remaining numbers. Continue with
the smallest number not already used, and pair it with any of the ﬁve numbers
43 2021 Solutions 183

remaining. Continuing, we see that the number of these partitions is .7∗5∗3∗1 =

105.
(b) Deﬁne

⎲
4
N (P ) =
. |xi − yi |2 , (43.1)
i=1

so that
⎛ ⎞ ⎛ 4 ⎞
⎲
4 ⎲
.N (P ) + 2V (P ) = (xi − yi ) + 2
2
xi yi
i=1 i=1

⎲
4 ⎲
8
= (xi2 + yi2 ) = k2
i=1 k=1

is constant (not depending on the choice of partition P ). So, .V (P ) is maximized

whenever .N (P ) is minimized; this occurs when the differences .|xi − yi | are as
small as possible. The partition

P1 = {{1, 2}, {3, 4}, {5, 6}, {7, 8}}

minimizes these differences and yields the value .V (P1 ) = 100. On the other
hand, maximizing .N (P ) is achieved by making the differences as large as
possible. The partition

P2 = {{1, 8}, {2, 7}, {3, 6}, {4, 5}}

does this, yielding a value .V (P2 ) = 60.

To check that .N (P2 ) = 84 is in fact the maximum, and that the above “greedy”
partition .P2 is the unique partition achieving it, consider the following Lemma:
Given any four real numbers in strictly increasing order: .x1 < x2 < x3 < x4 ,
deﬁne .N4 on partitions in analogy with (43.1) by
({ })
.N4 {xi , xj }, {xk , x𝓁 } = |xj − xi |2 + |x𝓁 − xk |2 ;

then .{{x1 , x4 }, {x2 , x3 }} is the unique partition maximizing .N4 .

Proof of Lemma Compare with the only other two partitions into pairs:

. N4 ({{x1 , x4 }, {x2 , x3 }}) − N4 ({{x1 , x2 }, {x3 , x4 }})

= (x4 − x1 )2 + (x3 − x2 )2 − (x1 − x2 )2 − (x3 − x4 )2
= 2x1 x2 + 2x3 x4 − 2x1 x4 − 2x2 x3. (43.2)
184 43 2021 Solutions

= 2(x4 − x2 )(x3 − x1 ) > 0,

N4 ({{x1 , x4 }, {x2 , x3 }}) − N4 ({{x1 , x3 }, {x2 , x4 }})
= (x4 − x1 )2 + (x3 − x2 )2 − (x3 − x1 )2 − (x4 − x2 )2
= 2x1 x3 + 2x2 x4 − 2x1 x4 − 2x2 x3 (43.3)
= 2(x4 − x3 )(x2 − x1 ) > 0.

The positivity of (43.2) and (43.3) is a special case of the “rearrangement

inequality.”
Suppose P is a partition of .{1, . . . , 8} where one of the pairs is .{k, 8} with
1 < k < 8. Then there is some other pair in the partition P of the form .{1, j }
.

with j distinct from 1, k, and 8. The Lemma applies (with either .1 < j < k < 8
or .1 < k < j < 8) and .N4 ({{k, 8}, {1, j }}) < N4 ({{1, 8}, {k, j }}). This shows
that

N (P ) = N ({{k, 8}, {1, j }, {x3 , y3 }, {x4 , y4 }})

< N ({{1, 8}, {k, j }, {x3 , y3 }, {x4 , y4 }}) ,

so .N(P ) is not a maximum. We can conclude that any partition maximizing N

must include the pair .{1, 8}.
Next, consider a partition P where one of the pairs is .{1, 8} and another is
.{k, 7} with .2 < k < 7. Then there is some other pair in the partition P of

the form .{2, j } with j distinct from 1, 2, k, 7, and 8. The Lemma applies and
.N4 ({{k, 7}, {2, j }}) < N4 ({{2, 7}, {k, j }}). This shows that

N (P ) = N ({{1, 8}, {k, 7}, {2, j }, {x4 , y4 }})

< N ({{1, 8}, {2, 7}, {j, k}, {x4 , y4 }}) ,

so .N(P ) is not a maximum. We can conclude that any partition maximizing N

must include the pairs .{1, 8} and .{2, 7}.
The only numbers remaining in any partition maximizing N are .{3, 4, 5, 6} and
the Lemma applies one last time, so that .P2 is the unique partition maximizing
N as claimed.

S2021-6A

(a) Let .a = 3, .b = 4, and .c = 7.

(b) The left-hand side is equivalent to .109a + 91c − 89b, which must equal 707 by
the conditions of the problem. Rewriting, .2(10a + b) = 707 − 89(a + c − b).
The left-hand side is even, which means .n = a + c − b must be odd. It is easy
to see that .n < 8, since .707 − 89 ∗ 8 < 0. Since a and b are single digits, the
largest that .(10a +b) can be is 99, which means only .n = 7 is a possibility. This
43 2021 Solutions 185

makes .(10a + b) = 42, making .a = 4 and .b = 2, and hence .c = 5. However,

this makes .109a + 91c − 89b = 713, not 707 as required.

S2021-6B Denoting the total by T and expanding give

T = abc + cab − bca = 109a − 89b + 91c = 101(a − b + c) + (8a + 12b − 10c).
.

The second term is even and satisﬁes

. − 90 ≤ 8a + 12b − 10c ≤ 180,

T − 180 ≤ 101(a − b + c) ≤ T + 90.

. (43.4)

For part (a), .T = 608 is even, and the only even solution of (43.4) is .a − b + c = 6,
so .8a + 12b − 10c = 2. For part (b), .T = 707 is odd, and the only odd solution
of (43.4) is .a − b + c = 7, so .8a + 12b − 10c = 0. These linear systems of equations
can be solved simultaneously by reducing the doubly augmented matrix
⎾ ⏋ ⎾ ⏋
1 −1 1 6 7 1 0 0.1 3.7 4.2
. → .
8 12 −10 2 0 0 1 −0.9 −2.3 −2.8

For part (a), the equations reduce to .10a + c = 37 and .10b − 9c = −23, with the
unique nonnegative single-digit solution vectors .a = 3, .b = 4, .c = 7. For part (b),
the equations reduce to .10a + c = 42 and .10b − 9c = −28, where .a = 4 and .c = 2
satisfy the ﬁrst equation, but then .10b − 18 = −28 has no nonnegative solution.

⎲
n
S2021-7 Let .p(x) = ak x k with .ak ≥ 0. Because .114 = 14641 > 2021, the
k=0
condition .p(11) = 2021 implies p is of the form .a3 x 3 + a2 x 2 + a1 x + a0 , and
because .2 · 113 = 2662 > 2021, .0 ≤ a3 ≤ 1. The condition that .p(1) = 21 means
the coefﬁcients must sum to 21, and hence each of the coefﬁcients cannot exceed
21. Considering the equation .a3 113 + a2 112 + a1 11 + a0 = 2021 modulo 11 gives
.a0 ≡ 8 (mod 11), so .a0 = 8 or .a0 = 19.

In the case .a0 = 19, the remaining coefﬁcients must add to 2, but none of the triples
(a3 , a2 , a1 ) = (0, 2, 0), .(0, 1, 1), .(0, 0, 2), .(1, 1, 0), or .(1, 0, 1) satisﬁes .p(11) =
.

2021.
In the case .a0 = 8 and .a3 = 0, .p(1) = 21 implies .a2 + a1 = 13 and .p(11) = 2021
implies .121a2 + 11a1 = 2013. The only solution of these two equations is .a2 = 17,
.a1 = −4, contradicting .a1 ≥ 0.
186 43 2021 Solutions

The only remaining case is .a0 = 8 and .a3 = 1. Then .p(1) = 21 implies .a2 + a1 =
12 and .p(11) = 2021 implies .121a2 + 11a1 = 682. The only solution of these two
equations is .a2 = 5, .a1 = 7.
The conclusion is that .p(x) = x 3 + 5x 2 + 7x + 8 and .p(10) = 1578.

Remark from the Editors If one additionally assumes that each coefﬁcient is a single
digit (.0 ≤ ak ≤ 9), then the condition that .p(11) = 2021 means that 2021 is
the base-10 value of a number with most signiﬁcant digit .an and one’s digit .a0 ,
expressed in base-11. The value .p(10) = 1578 can then be interpreted as the base-
11 representation of 2021.

S2021-8 The number 288 has a unique factorization in exactly 10 numbers from the
set .{1, 2, 3, 4, 5}, .288 = 14 23 32 41 50 . We can ﬁll in the different cells as follows:

• Because 288 cannot have any factors of five, we can fill in cells .(1, 1), .(5, 2),
.(3, 3), .(2, 4), and .(4, 5) with fives.

• Because the cage of 288 must have 4 ones as factors, we can ﬁll in cells .(1, 5),
.(2, 2), .(4, 3), and .(2, 4), and .(3, 4) with ones. Thus, cell .(5, 1) must be a one,

too.
• To get 3 factors of two in the 288 cage, cells .(1, 4) and .(1, 2) cannot be ﬁlled
with a two. It follows that cell .(1, 3) must be a two, and from there it follows that
cells .(3, 2) and .(4, 4) are also twos.
• The 288 cage is completed by inserting a four in cell .(1, 2) and threes in cells
.(4, 2) and .(1, 4).

• The remainder of the cells are forced, and it can be veriﬁed that the other cage’s
clue is also satisﬁed.

The solution is as follows.

5 4 2 3 1

2 1 4 5 3

3 2 5 1 4

4 3 1 2 5

1 5 3 4 2
Chapter 44
2022 Solutions

The 56th competition was held at Indiana Wesleyan University, after 2 years without
gathering in-person. Because of the transition back from remote to centralized
formats, there were fewer teams than normal participating, and the meeting itself
was held on a single day. While the competition was an in-person event, the scoring
and reporting was done virtually. Results were shared publicly within a few days
after the competition.
The 2022 problem set was prepared by Professor Colin McKinney at Wabash
College.
The 2022 problem statements begin on page 59 in Chap. 22.

S2022-1 This is a four-line locus problem, and so we should expect a conic section
as the resulting plane curve. The required equation is

. |x + 3||1 − x| = |y + 4||2 − y|.

There are two solutions, depending on the signs chosen when solving this equation
with absolute values. One is the hyperbola

(y + 1)2 − (x + 1)2 = 5,
.

and the other is the circle

(x + 1)2 + (y + 1)2 = 13.

Historical note from the problem author Compare Apollonius’ Conics III.54 [A]
and Descartes’ La Géométrie [De].

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 187
A. Coffman et al., The Indiana College Mathematics Competition (2001–2023),
Problem Books in Mathematics, https://doi.org/10.1007/978-3-031-62768-2_44
188 44 2022 Solutions

S2022-2 By the product rule, we need f and g to satisfy

f ' g ' = f ' g + g ' f.

For the sake of simplicity, choose .f (x) = x so that .f ' (x) = 1. Thus we require

g ' = g + xg ' .
.

Hence

g' 1
. = .
g 1−x

Integrating on some interval where .g /= 0, we see that

. ln |g(x)| = − ln |1 − x| + C,

A
so that .g(x) = for some constant A. Choosing .A = 1 just to find a specific,
1−x
1
a nonconstant example gives us functions .f (x) = x and .g(x) = . One can
1−x
easily check that these functions satisfy the requirements. Another pair of functions
that works is .e2x and .e2x .

Remark from the problem author This problem was inspired by ICMC P1966-4
[AFMC], which asks a similar question but with quotients.

S2022-3

(a) Standard differentiation techniques yield

e−1/x
f ' (x) = (
. )2 .
1 − e−1/x x 2

(b) Observe that, on .(0, ∞), f is an antiderivative of the integrand, by part (a), and
both f and .f ' extend continuously to .[0, ∞). Therefore, by the Fundamental
Theorem of Calculus,
⎰ 1 e−1/x
. dx = lim (f (1) − f (a))
0 x 2 (1 − e−1/x )2 a→0+

1
= .
e−1
44 2022 Solutions 189

Remark from the problem author This problem is a simplification of ICMC P1971-
3 [AFMC].

S2022-4 Let .B = I + A + A2 + · · · + Am−1 . Observe that .B(I − A) = I − Am = 0.

Since .A /= I , there is a vector .v→ such that .A→
v /= v→. Let .w
→ = (I − A)→
v = v→ − A→v.
Then .B w→ = B(I − A)→ v = 0. Thus, .w → is a nonzero vector in the null space of B,
and so .det B = 0.

Remark from the problem author This is the same as ICMC P1999-8 [AFMC].
This new solution was contributed by Bob Foote.

S2022-5

(a) .n? is a nonstandard notation for the nth triangular number, given by

(n)(n + 1)
n? = Tn =
. .
2
Thus .2022? = 1011 · 2023, which clearly has no zeros at the end of its decimal
representation.
(b) Divide 2022 by powers of 5 that are less than 2022, yielding .404+80+16+3 =
503. Thus, there are 503 zeros at the end of the decimal representation.

Remark from the Editors Part (b) is similar to ICMC P2012-2—see the solution
S2012-2 in Chap. 35.

S2022-6 No. Let G be the group, and suppose by way of contradiction that H and
K are proper subgroups with .G = H ∪ K. Because .H /⊆ K, we can take an element
.h ∈ H \ K. Similarly, let .k ∈ K \ H , and consider the element .hk ∈ G. Then

.hk ∈ H or .hk ∈ K. If .hk ∈ H , then .h

−1 hk ∈ H =⇒ k ∈ H , and if .hk ∈ K, then

.hkk
−1 ∈ K =⇒ h ∈ K; either case leads to a contradiction.

Remark from the problem author This is the same as ICMC P1972-7 [AFMC].

S2022-7 Note that .g(x) = g(y) =⇒ g m (x) = g m (y) =⇒ x = y. Therefore g

is injective. By a well-known result from real analysis (not proved here, but which
can be proved using the Intermediate Value Theorem), any injective real-valued
continuous function on an interval must be strictly monotone on the interval. If we
suppose that g is increasing on .[0, 1], then for .x ∈ [0, 1],

x > g(x) =⇒ g(x) > g 2 (x) =⇒ . . . =⇒ g m−1 (x) > g m (x),

hence .x > g m (x) = x, a contradiction. Similarly, if .x < g(x), we have .x <

g m (x) = x, another contradiction. Hence .g(x) = x. If on the other hand g is
190 44 2022 Solutions

decreasing, then .g 2 is increasing, and the above argument applies to show .g 2 (x) =
x. Hence, in either case, .g 2 (x) = x.

Remark from the problem author This problem and solution were both adapted
from ICMC P1983-9 [AFMC]. See also ICMC P1977-1.

S2022-8 The following example is constructed using the increasing sequence of

triangular numbers:

.(0, 1, 3, 6, 10, . . . , Ta , . . .),

defined by .Ta = 12 a(a + 1) or recursively by .T0 = 0 and .Ta+1 = Ta + a + 1. Define

.f : N × N → N by

f (m, n) = m + Tm+n−2 .
. (44.1)

To show f is surjective, consider .x ≥ 1, and let .a = max{y ∈ Z : Ty < x}, so that

Ta < x ≤ Ta+1 . Then .0 < x − Ta ≤ Ta+1 − Ta = a + 1, and one can verify that
.

f (x − Ta , a + 2 − (x − Ta )) = x.
.

To show f is injective, suppose there are two input pairs; label the pair with the
smaller or equal sum of components by .(m, n) and the other pair by .(m∗ , n∗ ), so
∗
.m + n ≤ m + n .
∗

We want to show that assuming the outputs are equal,

f (m, n) = f (m∗ , n∗ )
.

m + Tm+n−2 = m∗ + Tm∗ +n∗ −2 (44.2)

leads to the conclusion .m = m∗ and .n = n∗ .

Case 1. m+n < m∗ +n∗ . Then, using the increasing property of the .Ta sequence,
.

0 = Tm∗ +n∗ −2 − Tm+n−2 + m∗ − m

≥ Tm+n−1 − Tm+n−2 + m∗ − m
= m + n − 1 + m∗ − m
= m∗ + n − 1
≥ 1,

which is a contradiction.
44 2022 Solutions 191

Case 2. m + n = m∗ + n∗ . It follows immediately from (44.2) that .m = m∗ ; thus,

in this case, .n = n∗ .

Historical note from the problem author The .f (m, n) formula (44.1) was given by
Cantor, see [C].

Remark from the Editors Although there may be many formulas for bijections
.N × N → N, the Fueter–Pólya Theorem states that .f (m, n) and the expression
.f (n, m) = n+Tm+n−2 are the only quadratic polynomials defining such a bijection.

The appearance of Cantor’s construction in problems for students dates back to [M].
Chapter 45
2023 Solutions

The 57th competition was held at Indiana University Kokomo for the first time.
The 2023 problem set was prepared by Professor Emeritus Robert Foote at Wabash
College.
The 2023 problem statements begin on page 61 in Chap. 23.

The 2023 competition’s tied-for-first place team from Rose-Hulman. Left to Right: Professor Josh
Holden, Nathan Chen, Connor Lane, and Alexa Renner

S2023-1 We seek the maximum and minimum speeds of the particle, where speed
is the magnitude (absolute value) of velocity. Let

v(t) = x ' (t) = 7 + 6t − t 2 .

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 193
A. Coffman et al., The Indiana College Mathematics Competition (2001–2023),
Problem Books in Mathematics, https://doi.org/10.1007/978-3-031-62768-2_45
194 45 2023 Solutions

Looking for the max/min of .v(t) on the .[0, 10] interval, we have that .v ' (t) = 0 only
at .t = 3. Check that

v(0) = 7,
. v(3) = 16 and v(10) = −33.

The fastest it goes is 33 ft/s when .t = 10. Since .v(t) changes sign, the slowest it
goes is 0 ft/s, which happens when .t = 7.

S2023-2 Consider

. sin2 70◦ = sin2 50◦ + sin2 60◦ − 2 sin 50◦ sin 60◦ cos 70◦ .

This is the Law of Cosines for a triangle with sides .sin 50◦ , .sin 60◦ , and .sin 70◦ ,
assuming such a triangle exists. Since .50◦ + 60◦ + 70◦ = 180◦ , there are triangles
with angles .50◦ , .60◦ , and .70◦ . The Law of Sines states that

sin 50◦ sin 60◦ sin 70◦

. = = ,
a b c
where a, b, and c are the respective side lengths. This common ratio can be any
positive number—it simply scales the size of the triangle. When the ratio is 1, we
have that

a = sin 50◦ ,
. b = sin 60◦ , and c = sin 70◦

are the sides of a triangle.

S2023-3 Let .MX = {(x, y, z) ∈ [0, 1]3 | x ≥ y and x ≥ z}. This is the portion of
the cube on which x is the maximum. Then
⎰ 1⎰ 1⎰ 1 ⎰⎰⎰
f (x, y, z) dx dy dz = 3 x dx dy dz
0 0 0 MX
⎰ 1⎰ x ⎰ x ⎰ 1
. =3 x dy dz dx = 3 x 3 dx = 3/4.
0 0 0 0

Remark from the problem author This integral has a nice interpretation in prob-
ability. It is the expected value of .max{x, y, z} when x, y, and z are taken to be
independent, random variables on .[0, 1] with a uniform distribution.

S2023-4A Differentiating both sides yields this differential equation:

.f (x) + f (−x) = f ' (x) + f ' (−x).

45 2023 Solutions 195

The next step is not forced on us, but remembering that we are looking for some
function that works, it seems reasonable to consider functions that satisfy .f ' (t) =
f (t), namely .f (t) = aet . It is easy (and necessary) to check that this satisfies the
integral equation. The desired function of this form is .f (x) = 2023et .

S2023-4B The symmetries of the equation around .x = 0 may lead you to think
about even and odd functions. If so, you will quickly notice that the left side of the
equation is 0 if f is even, and the right side is 0 if f is odd. Thus, if a function
satisfies the equation, there must be some relation between its even and odd parts.
The simplest function that has both even and odd parts is .f (x) = ax + b. If you
plug this into both sides of the equation, it is a solution if and only if .a = b, so the
desired function of this form is .f (x) = 2023x + 2023.

S2023-4C More generally, any function .f (x) is the sum of its even and odd parts,
fe (x) = 12 (f (x) + f (−x)) and .fo (x) = 12 (f (x) − f (−x)). Then, assuming only
.

that f is integrable and .fe is continuous, the integral equation implies that .fo is
differentiable, and .fo' (x) = fe (x). This leads to some unusual functions satisfying
the equation, for example, choosing .fe = 2|x| + 2023 gives a nondifferentiable
solution .f (x) = x|x| + 2|x| + 2023x + 2023.

S2023-5A Proof by induction. Let .gn = fn+1 fn−1 − fn2 . We have .g2 = 2 · 1 − 1 =
(−1)2 for the base case. Suppose .gn−1 = (−1)n−1 . Then

gn = fn+1 fn−1 − fn2 = (fn + fn−1 )fn−1 − (fn−1 + fn−2 )fn

.
= fn−1
2
− fn fn−2 = −gn−1 = (−1)n .

S2023-5B Another proof also uses⎾induction and ⏋ a nice matrix equation that makes
fn+1 fn
the algebra a bit cleaner. Let .Fn = . We need to show .det Fn = (−1)n
fn fn−1
⎾ ⏋
11
for .n ≥ 2, where the base case .n = 2 is as in the previous solution. Let .E = ,
10
and note that

.Fn = EFn−1 . (45.1)

Suppose, for .n > 2, .det Fn−1 = (−1)n−1 . Then

. det Fn = det E det Fn−1 = (−1)(−1)n−1 = (−1)n .

Remark from the Editors A matrix product like (45.1) was considered in S2003-6
in Chap. 26.
196 45 2023 Solutions

S2023-6 Let L be the line parallel to .AB that is also tangent to .γ . Let Z be the
←→
intersection of L and .CH . Note that .ZX = H Y , since .ZX and .H Y are symmetric
across the diameter of .γ parallel to L and .AB (dotted line). Since Z is between C
and X (see the next paragraph), we have .CX > ZX = H Y .

To see that Z is between C and X, note that C and H are on opposite sides of L,
since L intersects the interior of .ΔABC but not .AB. On the other hand, X and H
←
→
are on the same side of L, since .γ is between L and .AB. Thus, C and X are on
opposite sides of L.

S2023-7 First, for each face-down card, consider the probability that it is black.
(This is a variation of the classic Monty Hall problem in which the black card is
the prize. This is worth looking up if you are unfamiliar with it.) Let .P (c is B)
denote the probability that card c is black. Initially (while the cards are all face
down), .P (c is B) = 1/4 for each c. Let W be the card you choose, let X be the
card your friend turns up, and let Y and Z be the other two cards. When your friend
turns up a card, it adds information about X, Y , and Z, but it adds no information
about W . In particular, .P (W is B) is still .1/4 and .P (X is B) = 0. By symmetry,
.P (Y is B) = P (Z is B). The four probabilities sum to 1, and so we have

P (W is B) = 1/4,
. P (Y is B) = 3/8, and P (Z is B) = 3/8.

Then (with R for red)

P (W is R) = 3/4,
. P (Y is R) = 5/8, and P (Z is R) = 5/8.

Remark from the problem author The claim that the revealed card adds no
information about W can be explained by drawing a probability tree for the problem.

S2023-8 Consider two beads that are about to collide. One moves to the right and
one to the left. After the collision, there is still one moving to the right and one
to the left. Thus, the numbers of right-moving beads and left-moving beads are
constant. After all the collisions there will still be n right-moving beads and k left-
moving beads, which answers part (a). If we ignore the bead colors, assume they are
negligibly small, and focus on the directions they move, the dynamics are the same if
45 2023 Solutions 197

we allow the beads to pass through each other instead of collide and rebound. Then
every right-moving bead passes through every left-moving bead. There are nk such
pass-through events (possibly some at simultaneous times, but spatially separated
because the right-moving beads maintain their initial pairwise distances from each
other), so in the original collision model, there are nk collisions, which answers
part (b).
Part III
More History of the ICMC
Chapter 46
Top Scoring Teams

The student participants in the 2015 ICMC, held at Taylor University

The following table contains the names of the winning teams for the period covered
by this volume. It includes some ties and, notably, reflects some early years in which
the section maintained the practice of not recognizing multiple high-scoring teams
from the same school.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 201
A. Coffman et al., The Indiana College Mathematics Competition (2001–2023),
Problem Books in Mathematics, https://doi.org/10.1007/978-3-031-62768-2_46
202 46 Top Scoring Teams

Year School Name

2001 1 Rose-Hulman Institute of Technology Andrew Chi
Ann Chi
2 Taylor University Shawn Alspaugh
David Aukerman
Noah DeLong
3 Wabash College Bogdan Ianev
Nigel Nunoo
Nick Roersma
3 Indiana University Greg Alexander
Craig Citro
Jason Grimblat
2002 1 IUPUI Bobby Ramsey
Muris Ridzal
Jon Landy
2 Wabash College Shiv Karunakaran
Bogdan Ianev
Daniel Smith
3 Taylor University David Aukerman
Shawn Alspaugh
Noah DeLong
3 Indiana University Craig Citro
Jason Grimblat
Seth Quackenbush
2003 1 Indiana—Purdue Fort Wayne Christian MacLeod
Jeff Wilkins
Kevin Chlebik
2 IUPUI Robert Walsman
Scott Pollum
Charles Tam
3 Wabash College Bogdan Ianev
Daniel Smith
Mick Roersma
2004 1 Rose-Hulman Institute of Technology Guy Srinivasan
Alex VanBrunt
Kellan Wampler
2 Taylor University Michael Anderson
Mike Assis
Shawn Burford
3 Ball State University
(continued)
46 Top Scoring Teams 203

Year School Name

2005 1 Purdue University Kyle Riggs
Brad Rodgers
Chris Willmore
2 Rose-Hulman Institute of Technology Stephen Dupal
Guy Srinivasan
Kellan Wampler
3 Rose-Hulman Institute of Technology Peter Fine
Curtis Katinas
Angela Smiley
2006 1 Rose-Hulman Institute of Technology Stephen Dupal
Robert Lemke-Oliver
Hari Ravindran
2 Purdue University Kyle Riggs
Brad Rodgers
Chris Willmore
3 Ball State University
2007 1 Rose-Hulman Institute of Technology Robert-Lemke-Oliver
Hari Ravindran
Amanda Rohde
2 Ball State University Neal Coleman
Wayne Drake
Nathan Pappas
3 Purdue University Jonathon Nistor
Nate Oriow
Prateek Tandon
3 Taylor University Jeremy Erickson
David Kasper
Matthew Russell
2008 1 Purdue University Noah Blach
Nate Oriow
Brad Rodgers
2 Rose-Hulman Institute of Technology Robert Lemke-Oliver
Hari Ravindran
3 Taylor University Jeremy Erickson
Matthew Russell
Jonathan Schrock
2009 1 Taylor University Jeremy Erickson
Matthew Russell
Joseph Seaborn
2 Indiana University Yun William Yu
Carlo Angiuli
Peter Lunts
(continued)
204 46 Top Scoring Teams

Year School Name

3 IUPUI Jeffrey Shen
Lyndon Ji
2010 1 Indiana University Carlo Angiuli
John Brown
Miles Dillon Edwards
2 Taylor University Jacob Erb
Matthew Russell
Joe Seaborn
3 Indiana Wesleyan University Tyler Carrico
Sam Johnston
Luke Nelson
2011 1 Indiana University John Brown
Kevin Carlson
Miles Dillon Edwards
2 IUPUI Kyle Firestone
Youkow Homma
William Karr
3 Taylor University David Ebert
Jacob Erb
Daniel Kasper
2012 1 Indiana University John Brown
Miles Dillon Edwards
Timothy Zakian
2 Wabash College Yifei Sun
Anh Tran
Tyler Koch
3 Indiana—Purdue Fort Wayne Valentin Bogun
Samuel Carolus
Altun Shukurlu
2013 1 Valparaiso University Timothy Goodrich
Michael Stuck
Ruyue (Julia) Yuan
2 Wabash College David Gunderman
Jia Qi
Xidian Sun
3 Taylor University Daniel Crane
Ethan Gegner
Josh Kiers
2014 1 Indiana University Tom Dauer
Jonathan Hawkins
Max Zhou
1 Taylor University Ethan Gegner
Josh Kiers
Claire Spychalla
(continued)
46 Top Scoring Teams 205

Year School Name

3 Indiana—Purdue Fort Wayne Vreneli Brenneman
Guchen Liu
Altun Shukurlu
3 Taylor University Sam Judge
Jordan Melendez
Justin Wydra
2015 1 Taylor University Daniel Crane
Josh Kiers
Claire Spychalla
2 Rose-Hulman Institute of Technology Ian Ludden
Christian Schulz
Lujia Zhang
3 Indiana—Purdue Fort Wayne Vreneli Brenneman
Altun Shukurlu
Sofia Sorokina
2016 1 Indiana University Tom Dauer
Baptiste Dejean
Max Zhou
2 Indiana Wesleyan University Jeremy David
Chriss Foss
Luke Hayden
3 Rose-Hulman Institute of Technology Jacob Hiance
Adam Michael
2017 1 Indiana University Ben Briggs
Andrew Henderson
Baptiste Dejean
2 Wabash College Tu Nguyen
Ngoc Tran
Thang Nguyen
3 Earlham College Vuong Khuat
Hiep Nguyen
Son Tran
2018 1 Indiana University Anthony Coniglio
Baptiste Dejean
Nathanial Lowry
2 Purdue University Israel Baker
Kevin LaMaster
Junyao Wu
3 University of Notre Dame Kyle Duffy
Matt Schoenbauer
Caitlyn Booms
3 Indiana—Purdue Fort Wayne Vincent Rivera
Giang Le
Duy Anh Do
(continued)
206 46 Top Scoring Teams

Year School Name

2019 1 Rose-Hulman Institute of Technology Zixin Fan
Ruizin Feng
Jiashen Hu
2 Indiana University Anthony Coniglio
Mathanial Lowry
Tiancheng Xu
3 Taylor University Drew Anderson
Jordan Crawford
Alexander McFarland
2021 1 Purdue University Ethan Brady
Sterling Saint-Rain
Chiara Travesset
2 Rose-Hulman Institute of Technology Reed Phillips
Bohdan Vakhitov
Max Li
3 Valparaiso University Eric Burkholder
Spencer Gannon
Marcus Engstrom
3 Purdue University Max Martinez
Kyle Kraft
Colton Griffith
2022 1 Rose-Hulman Institute of Technology Reed Phillips
Ben Lyons
2 Purdue University Kyle Kraft
Max Martinez
Colton Griffen
2 IUPUI Sara Hiatt
Croix Gyurek
Clayton Kelley
2023 1 Notre Dame Samuel Dekleva
Molly MacDonald
Philip Templeman
1 Rose-Hulman Institute of Technology Nathan Chen
Connor Lane
Alexa Renner
3 Notre Dame Andrew Brooks
Gavin Dooley
Zachary Joseph
Chapter 47
Updates to A Friendly Mathematics
Competition

Some Errata for the 1966–2000 Problem Book

Page numbers for these corrections refer to the publication [AFMC].

• p. 15: The team member’s name is Tom Sellke; he also appears on p. 16.
• p. 30: The word in line 7 of P1987-3 should be “falling.”
• p. 30: The distance formula in P1987-5 should be
⎛ ⎞1/2
|XY | = (y1 − x1 )2 + (y2 − x2 )2 + (y3 − x3 )2 + (y4 − x4 )2
. .

• p. 33: The notation should indicate an inﬁnite sequence

{x1 , x2 , . . . , xk , . . .}.
.

• p. 38: The team member’s name is Mike Bolt.

• p. 42: The symbol in P1996-5 should be “degrees”: ./ A = 90◦ .
• p. 60: The symbol in paragraph 2, line 2, of S1967-8 should be a bold .Q.
• p. 62: The symbol in line 4 of S1968-3 should have a prime for the derivative:
'
.tf (tx). See also comments in section “New Solutions for Old Problems”.

• p. 87: The expression in line 2 of S1976-4 should have an equal sign: .f ' (1) =
3 ln 3 − 4 < 3.3 − 4 < 0.
• p. 88: The symbol in the last line of S1976-6 should be a strict inequality: .f (x) <
0.
• p. 126: The number in line 3 of S1987-6 should be .2π(4000)/4.
• p. 131: The equation in S1989-3 should be .2(f + c) = a + d + g.
• p. 167: The symbol in S1998-4 in the last sentence on this page should be a
bold .F.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 207
A. Coffman et al., The Indiana College Mathematics Competition (2001–2023),
Problem Books in Mathematics, https://doi.org/10.1007/978-3-031-62768-2_47
208 47 Updates to A Friendly Mathematics Competition

• p. 172: In S1999-3, some y terms should be .y 2 , so that line (ii) is

√ /
5x 2 + 2xy + y 2 ≤ 2x 2 + 2 2x x 2 + y 2 + x 2 + y 2
.

and line (iv) is

x 2 + 2xy + y 2 ≤ 2(x 2 + y 2 ).
.

New Solutions for Old Problems

P1968-3 Let .f : Rn → Rn be a differentiable function such that .f (tx) = tf (x)

for .x ∈ Rn and .t > 0. Show that f is linear.
The solution appearing in [AFMC] is short and elementary but seems to use an
unstated, and unnecessary, assumption that f is continuously differentiable.
Here we state and prove a more general result.

Theorem Let f : Rn → Rm be a function which is differentiable at the origin, and

such that f (t x→) = tf (→
x ) for x→ ∈ Rn and t > 0. Then f is linear.

Proof By the deﬁnition of multivariable differentiability, there is some constant

m × n matrix M so that

f (0→ + h) → − M · h→
→ − f (0)
→
. lim = 0.
→ 0→
h→ →
‖h‖

We also have, by hypothesis,

→ = f (2 · 0)
f (0)
. → = 2 · f (0)
→ =⇒ f (0)
→ = 0.
→

So, the differentiability property is

→ − M · h→
f (h)
. lim →
= 0.
→ 0→
h→ →
‖h‖

→ and let ϵ be an arbitrary positive number. Considering

Fix x→ /= 0, ϵ
> 0, there is
‖→
x‖
a corresponding δ > 0 so that if 0 < ‖h‖→ < δ, then

→ − M · h‖
‖f (h) → ϵ
. < .
→
‖h‖ ‖→
x‖
New Solutions for Old Problems 209

Pick t > 0 so that ‖t x→‖ < δ. Then

‖f (t x→) − M · (t x→)‖ ϵ
. <
‖t x→‖ ‖→
x‖
1
=⇒ ‖tf (→
x ) − tM · x→‖ < ϵ
t
=⇒ ‖f (→ x ) − M · x→‖ < ϵ,

which, since ϵ and x→ were arbitrary, implies f (→ x ) = M · x→ for any x→ /= 0. → The

x ) is equal to the linear expression M · x→ for all x→ ∈ R .
conclusion is that f (→ n

P1969-6 Assume that f has a continuous second derivative, that a < b, that
f (a) = f (b) = 0, and that |f '' (x)| ≤ M on a ≤ x ≤ b. Prove that
|⎰ |
| b | M
.| f (x)dx || ≤ (b − a)3 .
| 12
a

The solution appearing in [AFMC] identifies this inequality as a special case of the
error estimate for the Trapezoidal Rule for integrals and then refers the reader to the
literature to find a proof. Here we sketch such a proof for the claimed special case.
For t in an open interval containing [a, b], define this difference of areas:
⎰ t 1
g(t) =
. f (x)dx − (t − a)f (t),
a 2
⎰b
so that g(a) = 0 and g(b) = a f (x)dx (using f (b) = 0). g is differentiable, with

1 1
g ' (t) =
. f (t) − (t − a)f ' (t),
2 2

and g ' (a) = 0 (using f (a) = 0). The second derivative is continuous by hypothesis

1
.g '' (t) = − (t − a)f '' (t).
2
Using the Fundamental Theorem of Calculus,
|⎰ | |⎰ |
| b | | b |
| f (x)dx || = |g(b) − g(a)| = || g ' (x)dx ||
.
|
a a
|⎰ ⎛ ⎰ ⎞ |
| b x |
= || g ' (a) + g '' (y)dy dx ||
a a
210 47 Updates to A Friendly Mathematics Competition

⎰ ⎰ ⎰ ⎰ | |
b x | '' | b x | 1 |
≤ |g (y)| dydx = |− (y − a)f '' (y)| dydx
| 2 |
a a a a
⎰ b ⎰ x ⎰ b
1 M M
≤ (y − a)Mdydx = (x − a)2 dx = (b − a)3 .
a a 2 4 a 12

P1982-5 A real-valued function f of a real variable is said to satisfy a Hölder

condition with exponent α if there is a constant c such that |f (x)−f (y)| ≤ c|x−y|α
for all x, y. Wherever these functions are used, α is restricted to be ≤ 1. Can you
explain why?
The solution appearing in [AFMC] states that if α ≤ 1, then f is uniformly
continuous on its domain. This is true for f on an interval where the condition
is satisﬁed with 0 < α ≤ 1, for example, f (x) = |x|α on the interval [−1, 1], but it
seems to be an incomplete answer to the question.
Here we state and prove the following result.

Theorem If f (x) has the property that there exist c ≥ 0 and α > 1 such that
|f (x) − f (y)| ≤ c|x − y|α for all x, y in an open interval (a, b), then f is constant
on (a, b).

Proof Fix y ∈ (a, b), and consider this expression depending on x ∈ (a, b), x /= y:
| |
| f (x) − f (y) | |f (x) − f (y)|
| |=
.
| x−y | |x − y|
c|x − y|α
≤ = c|x − y|α−1 .
|x − y|

Because α − 1 > 0, the Squeeze Theorem for limits applies to show that the limit
exists:
f (x) − f (y)
f ' (y) = lim
. = lim c|x − y|α−1 = 0.
x→y x−y x→y

Since the derivative is zero at every point y in the interval, the function is constant
(as in P2016-6).

P1986-3 Evaluate i i .
The solution appearing in [AFMC] gives a correct answer, i i = e−π/2 . However,
complex exponentiation is multivalued, so there are inﬁnitely many other answers
(which in this case are all real numbers): For any integer n,

i i = (e(π i/2)+2π in )i = e−π/2−2π n .

.
New Solutions for Old Problems 211

This sequence can also be explained in terms of the multivalued property of the
complex logarithm:

i i = ei log(i) = ei((π i/2)+2π in) .

P1992-6c For an arbitrary 2 × 2 matrix A, what is the maximum number of integral

matrices B that can satisfy B 2 = A?
⎾ n ⏋
2 0
The solution appearing in [AFMC] gives the example A = and the 2(n +
0 2n
⎾ ⏋
0 2i
1) matrices B = ± n−i , for 0 ≤ i ≤ n, concluding that there is no maximum.
2 0
⎾ ⏋
00
Another example, where there are inﬁnitely many solutions, is A = and the
00
⎾ ⏋
0i
matrices B = , for any integer i.
00
Appendix

Location Index

The following educational institutions in Indiana appear in this book (or the earlier
volume [AFMC], for 1966–2000) as host venues for Indiana MAA meetings, or as
the home field for winning teams or problem writers.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 213
A. Coffman et al., The Indiana College Mathematics Competition (2001–2023),
Problem Books in Mathematics, https://doi.org/10.1007/978-3-031-62768-2
214 Appendix

1. Anderson University
2. Ball State University
3. Butler University
4. DePauw University
5. Earlham College
6. Franklin College
7. Goshen College
8. Hanover College
9. Huntington University
10. Indiana State University
11. Indiana University Bloomington
12. Indiana University East
13. Indiana University Kokomo
14. Indiana University Northwest
15. Indiana University - Purdue University Fort Wayne
16. Indiana University - Purdue University Indianapolis
17. Indiana Wesleyan University
18. Manchester College
19. Marian College
20. Purdue University Calumet
21. Purdue University North Central
22. Purdue University West Lafayette
23. Rose-Hulman Institute of Technology
24. St. Joseph’s College
25. Saint Mary-of-the-Woods College
26. Saint Mary’s College
27. Taylor University
28. Trine University (formerly Tri-State University)
29. University of Evansville
30. University of Indianapolis
31. University of Notre Dame
32. University of Southern Indiana
33. Valparaiso University
34. Wabash College
Appendix 215

26 31
21 28
20 14 7
33

15
18
24 9

17
22 13 27

2
1
34
12
3
19 16 5
30
4
25
23 6
10

32
29
216 Appendix

References and Photo Credits

[A] Apollonius, On Conic Sections, translated by R. Catesby Taliaferro. University of

Chicago Great Books 11 (Encyclopædia Britannica, Inc., 1952)
[C] G. Cantor, Contributions to the Founding of the Theory of Transfinite Numbers,
translated by P. Jourdain (Dover Publications, Inc., 1955). Reprint of 1915 publication
by Open Court Pub. Co., Chicago
[CZ] A. Coffman, Y. Zhang, Vector fields with continuous curl but discontinuous partial
derivatives. Am. Math. Mon. (10) 127, 911–917 (2020)
[Da] H. Davis, The first meeting of the Indiana Section. Am. Math. Mon. (1) 32, 37 (1925)
[De] R. Descartes, The Geometry, translated by D. Smith, M. Latham, with a facsimile of
the first edition, 1637 (Dover Publications, Inc., 1954), reprint of 1925 publication by
Open Court Pub. Co., Chicago
[ME22 ] Margaret Edson, private communication with RG, November 2022
[ME23 ] Margaret Edson, private communication with RG, June 2023
[MA] Mary (Edson) Ales, private communication with RG, April 2023
[GKL] G. Gilbert, M. Krusemeyer, L. Larson, The Wohascum County Problem Book.
Dolciani Mathematical Expositions 14 (Mathematical Association of America, 1993)
[AFMC] R. Gillman (ed.), A Friendly Mathematics Competition. 35 Years of Teamwork in
Indiana. MAA Problem Books Series 8 (MAA Press: An Imprint of the American
Mathematical Society, 2003)
[G] H. Grossman, The twelve-coin problem. Scr. Math. 11, 360–361 (1945)
[GN] R. Guy, R. Nowakowski, Coin-weighing problems. Am. Math. Mon. (2) 102, 164–
167 (1995)
[HWr] G. Hardy, E. Wright, An Introduction to the Theory of Numbers, 5th edn. (Oxford,
1979)
[HWi] K. Hardy, K. Williams, The Green Book of Mathematical Problems (Dover Publica-
tions, Inc., 1997). Reprint of 1985 publication by Integer Press, Ottawa
[K1 ] J. Kürschák, G. Hajós, G. Neukomm, J. Surányi (eds.), Hungarian Problem Book I.
Based on the Eötvös Competitions, 1894 − 1905, translated by E. Rapaport. New
Mathematical Library 11 (Random House, 1963)
[K2 ] J. Kürschák, G. Hajós, G. Neukomm, J. Surányi (eds.), Hungarian Problem Book II.
Based on the Eötvös Competitions, 1906 − 1928, translated by E. Rapaport. New
Mathematical Library 12 (Random House, 1963)
[M] G. Mathews, Theory of Numbers. Part I (Deighton, Bell & Co., Cambridge, 1892)
[M1 ] P. Mielke, Improper Integrals in Abstract Spaces, Purdue University Ph.D. disserta-
tion, 1951
[M2 ] P. Mielke, Rational points on the number line. Math. Teach. (6) 63, 475–479 (1970)
[M3 ] P. Mielke, A tiling of the plane with triangles. Two-Year Coll. Math. J. (5) 14, 377–
381 (1983)
[PFW] Photograph courtesy J. Whitcraft, Purdue University Fort Wayne Communications
and Marketing
[RL] R. Rusczyk, S. Lehoczky, The Art of Problem Solving. Volume 2: and Beyond, 7th
edition with separate Solutions Manual (AoPS Inc., 2013)
[S] D. Struik (ed.), A Source Book in Mathematics, 1200 − 1800 (Harvard University
Press, 1969)
[TU] Photograph courtesy J. Case, Taylor University Department of Mathematics
[WCLib] Photographs of Paul Mielke, P-21, Wabash Photograph Collection, Robert T. Ramsay
Jr. Archival Center, Lilly Library, Wabash College, Crawfordsville
[WCNews] P. Mielke, Jr. et al., Memorial Service for Professor Emeritus Paul Mielke ’42,
Wabash College web site news page (2008). https://www.wabash.edu/news/story/
6013
Appendix 217

[W] S. Weintraub, An observation on average velocity. Am. Math. Mon. (4) 130, 384
(2023)
[WHS] Peter Edson Papers, Archives of the Wisconsin Historical Society, Madison. http://
digital.library.wisc.edu/1711.dl/wiarchives.uw-whs-mss00730
Index

A P2019-5, 52
Algebra P2021-1, 57
abstract P2022-4, 59
P2003-1, 7 S2013-6B, 134
P2005-4, 11 S2021-6B, 185
P2008-1, 17 S2023-5B, 195
P2015-7, 37 polynomials
P2016-2, 39 P2007-1, 15
groups P2012-3, 27
P2002-5, 6 P2021-2, 57
P2007-6, 16 P2021-7, 58
P2009-3, 19 S2008-7A, 102
P2012-5, 27 S2009-4, 104
P2013-4, 29 S2016-5, 154
P2013-5, 29 rings
P2014-7, 32 P2002-5, 6
P2016-7, 40 P2006-5, 14
P2019-7, 52 S2005-5A, 87
P2022-6, 60
matrix B
P1992-6c, 211 Binomial Theorem
P2003-4, 7 S2005-8, 88
P2003-6, 8 S2012-1, 121
P2004-3, 9 S2014-8, 142
P2005-5, 11 S2015-2A, 146
P2007-5, 16
P2008-7, 18 C
P2010-4, 21 Calculus
P2011-8, 25 differentiation
P2012-4, 27 P2015-8, 37
P2014-3, 31 P2016-6, 40
P2016-5, 40 P2021-2, 57
P2017-6, 47 P2023-1, 61
P2018-4, 49 S2009-6B, 105

© The Editor(s) (if applicable) and The Author(s), under exclusive license to 219
Springer Nature Switzerland AG 2024
A. Coffman et al., The Indiana College Mathematics Competition (2001–2023),
Problem Books in Mathematics, https://doi.org/10.1007/978-3-031-62768-2
220 Index

Calculus (cont.) S2015-7, 149

first semester Cayley graph
P2001-2, 3 S2019-7, 178
P2004-6, 10 Cayley-Hamilton Theorem
P2006-1, 13 S2008-7A, 102
P2006-4, 13 Combinatorics
infinite series P2001-3, 3
P2001-6, 3 P2002-4, 5
P2002-2, 5 P2010-1, 21
P2003-5, 8 P2013-3, 29
P2004-2, 9 P2015-2, 35
P2006-2, 13 P2016-8, 41
P2009-2, 19 P2017-7, 47
P2017-2, 46 P2018-8, 50
S2010-5, 110 P2019-3, 51
integration P2021-5, 57
P1969-6, 209 P2023-8, 62
P2005-1, 11 Comparison Test
P2007-3, 15 S2001-6, 68
P2012-7, 27 S2002-2, 72
P2022-3, 59 S2003-5, 78
P2023-4, 61 S2017-2, 161
S2006-3, 92 Complex numbers
limits P1986-3, 210
P2005-6, 11 P2011-8, 25
P2007-3, 15 S2001-4A, 67
P2010-5, 22
P2011-3, 24
P2011-6, 25 D
P2012-7, 27 Differential equations
P2014-1, 31 P2012-8, 28
P2014-2, 31 P2022-2, 59
multivariable S2023-4A, 194
P1968-3, 208 Discrete math
P2008-4, 17 P2011-1, 23
P2011-3, 24 P2011-5, 25
P2012-8, 28 P2015-4, 35
P2023-3, 61
S2019-8, 179 E
sequences Epsilon-delta
P2004-4, 9 P1968-3, 208
P2005-3, 11 S2009-6A, 104
P2008-5, 17 S2010-7, 111
P2013-1, 29 S2012-7, 126
P2017-5, 47 S2014-2, 138
P2018-5, 49 Euler’s Formula
P2019-6, 52 P2013-8, 30
volumes Euler’s Identity
P2008-4, 17 S2001-4A, 67
P2011-2, 23
P2019-8, 53
S2014-3A, 138 F
S2015-3, 147 Fermat’s Little Theorem
Cauchy’s Theorem S2005-8, 88
Index 221

Fibonacci sequence S2010-3, 109

P2003-6, 8 S2011-6, 117
P2009-2, 19 S2012-6, 125
P2013-6, 29 S2013-1, 131
P2016-5, 40 S2013-6A, 133
P2023-5, 61 S2014-2, 138
Four Color Theorem S2015-1, 145
S2015-6, 149 S2015-5, 148
S2016-1, 151
S2016-2A, 151
G S2016-2B, 152
Games S2016-5, 154
P2003-2, 7 S2017-6, 164
P2010-3, 21 S2019-5, 177
P2011-1, 23 S2023-5A, 195
P2011-5, 25 Inequalities
Geometry arithmetic-geometric mean
analytic S2014-3B, 139
P2001-5, 3 S2014-3C, 140
P2011-2, 23 rearrangement
P2019-2, 51 S2021-5, 182
P2019-8, 53 triangle
P2021-4, 57 S2011-6, 117
P2022-1, 59 Intermediate Value Theorem
Euclidean S2006-4, 92
P2002-6, 6 S2014-4, 140
P2003-3, 7 S2022-7, 189
P2004-5, 10
P2005-2, 11
P2005-7, 12 L
P2006-7, 14 Law of Cosines
P2014-5, 32 S2003-3, 76
P2015-3, 35 S2006-7, 93
P2016-3, 39 S2018-6B, 169
P2017-1, 45 S2023-2, 194
P2018-6, 50 Law of Sines
P2023-6, 62 S2003-3, 76
Graph theory S2014-5, 140
P2013-8, 30 S2023-2, 194
P2015-6, 36 L’Hôpital’s Rule
Green’s Theorem S2014-1, 138
S2012-8, 128 Linear algebra, see Algebra–matrix
Logic puzzle
P2006-8, 14
H P2015-4, 35
Heine-Borel Theorem P2021-8, 58
S2013-2, 132

I M
Induction Mean Value Theorem
S2003-2, 76 S2011-6, 117
S2003-6, 79 S2015-8, 150
S2004-4, 82 S2016-6, 156
222 Index

Monotone Convergence Theorem R

S2013-1, 131 Rational Root Theorem
S2019-6, 177 S2001-4A, 67
S2012-3, 122
Ratio Test
N S2004-2, 81
Number theory Real analysis
analytic P1982-5, 210
P2001-4, 3 P2002-1, 5
P2011-4, 24 P2009-6, 20
P2012-6, 27 P2010-7, 22
elementary P2012-6, 27
P2001-1, 3 P2013-7, 30
P2002-3, 5 P2014-4, 32
P2005-8, 12 P2015-5, 36
P2006-6, 14 P2017-8, 47
P2007-2, 15 P2018-2, 49
P2007-4, 15 P2018-3, 49
P2008-6, 18 P2019-1, 51
P2009-4, 20 P2022-7, 60
P2009-5, 20 Rolle’s Theorem
P2010-2, 21 S2001-2, 66
P2012-1, 27 S2015-5, 148
P2012-2, 27
P2014-8, 33
P2015-1, 35 S
P2016-1, 39 Set theory
P2017-4, 46 P2004-1, 9
P2021-3, 57 P2009-7, 20
P2021-6, 58 P2011-7, 25
P2022-5, 60 P2014-6, 32
Numerical analysis P2022-8, 60
P1969-6, 209 Social science
P2002-4, 5
P2003-2, 7
P P2014-6, 32
Pigeonhole principle P2016-8, 41
S2002-4, 72 P2018-8, 50
S2013-3, 132 P2019-3, 51
S2016-7, 156 S2015-2B, 146
S2018-8, 170 Squeeze Theorem
Probability P1982-5, 210
P2003-2, 7 S2005-3, 86
P2006-3, 13 S2015-5, 148
P2007-4, 15
P2008-2, 17 T
P2009-1, 19 Telescoping sum
P2010-6, 22 S2004-2, 81
P2017-3, 46 S2006-2, 91
P2018-7, 50 S2008-5, 100
P2019-4, 52 S2011-5, 115
P2023-7, 62 S2018-2, 167
S2023-3, 194 S2018-5, 168
Index 223

Topology S2003-3B, 78
P2013-2, 29
P2016-4, 39
Triangular numbers V
S2001-1, 65 Vectors
S2009-7A, 106 S2001-5, 68
S2017-5, 163 S2014-3A, 138
S2022-5, 189 S2021-4, 182
S2022-8, 190
Trigonometry
P2008-3, 17 W
P2018-1, 49 Wilson’s Theorem
P2023-2, 61 S2009-5, 104
S2001-4B, 67 S2010-2, 109

TRX Official Trainer Guide Filled - Password - Removed PDF
100% (10)
TRX Official Trainer Guide Filled - Password - Removed PDF
52 pages
Defamation Claim
No ratings yet
Defamation Claim
13 pages
Kazakhstan 2021
No ratings yet
Kazakhstan 2021
45 pages
Regional Mathematical Olympiad (Rmo) - 2019: Test Paper With Solution
100% (2)
Regional Mathematical Olympiad (Rmo) - 2019: Test Paper With Solution
6 pages
MAA The Contest Problem Book 3 (III) PDF
100% (2)
MAA The Contest Problem Book 3 (III) PDF
191 pages
Newyork
0% (1)
Newyork
250 pages
CompendiumJBMO Removed
No ratings yet
CompendiumJBMO Removed
357 pages
IGO Problems + Solutions 2014-2022
100% (1)
IGO Problems + Solutions 2014-2022
312 pages
MR Year 2006 Problem Column PDF
100% (1)
MR Year 2006 Problem Column PDF
218 pages
Turkey: National Olympiad First Round 1998
No ratings yet
Turkey: National Olympiad First Round 1998
5 pages
The Contest Probelm Book IV
No ratings yet
The Contest Probelm Book IV
201 pages
Diophant Lecture Notes PDF
No ratings yet
Diophant Lecture Notes PDF
65 pages
VN Booklet 2021
100% (1)
VN Booklet 2021
42 pages
FileGocHOMOMoi
100% (3)
FileGocHOMOMoi
235 pages
2010 Arml
No ratings yet
2010 Arml
45 pages
Saudi Booklet 2021
100% (1)
Saudi Booklet 2021
38 pages
Komal Advanced Problems
100% (1)
Komal Advanced Problems
16 pages
Shailesh A Shirali Arithmetic and Algebra Ramanujan Mathematical Society Little Mathematical Treasures PRMO RMO INMO Math Olympiad IMO (PDFDrive)
No ratings yet
Shailesh A Shirali Arithmetic and Algebra Ramanujan Mathematical Society Little Mathematical Treasures PRMO RMO INMO Math Olympiad IMO (PDFDrive)
356 pages
Classical Inequalities - Ivan Matic
No ratings yet
Classical Inequalities - Ivan Matic
24 pages
HOMCproblem&solution 2006 2016 PDF
100% (1)
HOMCproblem&solution 2006 2016 PDF
77 pages
Hojoo Lee & Peter Vandendriessche - Problems in Elementary Number Theory
100% (1)
Hojoo Lee & Peter Vandendriessche - Problems in Elementary Number Theory
88 pages
Problems: Intermediate Level
No ratings yet
Problems: Intermediate Level
17 pages
Geometric Inequalities (Akadii Slinko)
No ratings yet
Geometric Inequalities (Akadii Slinko)
24 pages
Competition Corner (Tay Tiong Seng)
No ratings yet
Competition Corner (Tay Tiong Seng)
20 pages
Math145fall09 Book PDF
No ratings yet
Math145fall09 Book PDF
146 pages
Principles and Techniques in Combinatorics Ebooknbsped 981436567x 9789814365673
No ratings yet
Principles and Techniques in Combinatorics Ebooknbsped 981436567x 9789814365673
368 pages
MCL 16.e
100% (1)
MCL 16.e
169 pages
RMC 2014
100% (5)
RMC 2014
104 pages
Nordic Mathematical Contest Problems and Solutions, 1987-2011 Problems
No ratings yet
Nordic Mathematical Contest Problems and Solutions, 1987-2011 Problems
68 pages
IMOTC Number Theory
No ratings yet
IMOTC Number Theory
28 pages
EGMO 2012-19 EN With Solutions PDF
100% (1)
EGMO 2012-19 EN With Solutions PDF
151 pages
Bulgarian Mathematical Olympiads (Third and Fourth Rounds From 1995 To 2000) (More Than 300 Problems With Solutions) - 260p
67% (3)
Bulgarian Mathematical Olympiads (Third and Fourth Rounds From 1995 To 2000) (More Than 300 Problems With Solutions) - 260p
260 pages
Problem Shortlist: Ego Loss May Occur
No ratings yet
Problem Shortlist: Ego Loss May Occur
59 pages
MCL 12.e
No ratings yet
MCL 12.e
194 pages
Geometric Inequalities Marathon - The First 100 Problems and Solutions
100% (1)
Geometric Inequalities Marathon - The First 100 Problems and Solutions
40 pages
Book Homc2019 Đã Chuyển Đổi
No ratings yet
Book Homc2019 Đã Chuyển Đổi
208 pages
Malaysian IMO 2007-2008 Booklet
No ratings yet
Malaysian IMO 2007-2008 Booklet
8 pages
Austrict 2021
No ratings yet
Austrict 2021
20 pages
Bmo SL 2022
No ratings yet
Bmo SL 2022
45 pages
HRV ABooklet 2015
100% (1)
HRV ABooklet 2015
38 pages
Hanoi Open Mathematical Olympiad 2006-2009
No ratings yet
Hanoi Open Mathematical Olympiad 2006-2009
16 pages
INMOTC 2020 Problem Set
No ratings yet
INMOTC 2020 Problem Set
7 pages
Saudi Booklet 2019 Final
100% (2)
Saudi Booklet 2019 Final
63 pages
Saudi Arabia IMO Booklet 2024 (Final)
No ratings yet
Saudi Arabia IMO Booklet 2024 (Final)
61 pages
2018 Bmo Shortlist
No ratings yet
2018 Bmo Shortlist
36 pages
RMC2006
100% (1)
RMC2006
97 pages
Armenian Math Olympiad
No ratings yet
Armenian Math Olympiad
1 page
ANGLE BISECTORS - San Jose Math Circle
No ratings yet
ANGLE BISECTORS - San Jose Math Circle
20 pages
The Solution of Equations in Integers
From Everand
The Solution of Equations in Integers
A. O. Gelfond
No ratings yet
Combinatorics Decoded
From Everand
Combinatorics Decoded
A K Pandey
5/5 (1)
Induction in Geometry
From Everand
Induction in Geometry
L.I. Golovina
No ratings yet
Challenging Math Problems
From Everand
Challenging Math Problems
Terry Stickels
No ratings yet
A Friendly Mathematics Competition - 35 Years of Teamwork in Indiana
89% (9)
A Friendly Mathematics Competition - 35 Years of Teamwork in Indiana
197 pages
A Friendly Mathematics Competition - 35 Years of Teamwork in Indiana (Gillman) 0883858088 PDF
100% (3)
A Friendly Mathematics Competition - 35 Years of Teamwork in Indiana (Gillman) 0883858088 PDF
197 pages
The Contest Problem Book III
100% (1)
The Contest Problem Book III
191 pages
MAA - The Contest Problem Book III, 1966-1972 (Annual High School Contests) - Charles T. Salkind, James M. Earl - (New Mathematical Library Vol 25) - MAA, 1973 - 186p
No ratings yet
MAA - The Contest Problem Book III, 1966-1972 (Annual High School Contests) - Charles T. Salkind, James M. Earl - (New Mathematical Library Vol 25) - MAA, 1973 - 186p
191 pages
Contest Problem Book III Annual High School Contest 1966-1972 - Salkind and Earl
No ratings yet
Contest Problem Book III Annual High School Contest 1966-1972 - Salkind and Earl
191 pages
Benefits of A Math Club Association
No ratings yet
Benefits of A Math Club Association
9 pages
The NYC Contest Problem Book
91% (11)
The NYC Contest Problem Book
250 pages
Judita Cofman - What To Solve - Problems and Suggestions For Young Mathematicians (1990)
No ratings yet
Judita Cofman - What To Solve - Problems and Suggestions For Young Mathematicians (1990)
264 pages
In What Sense Is It True To Claim That Mathematics Is Culture-Free?
No ratings yet
In What Sense Is It True To Claim That Mathematics Is Culture-Free?
5 pages
MAA The Contest Problem Book 2 (II) PDF
No ratings yet
MAA The Contest Problem Book 2 (II) PDF
120 pages
Error TrapezoidalRule
No ratings yet
Error TrapezoidalRule
3 pages
Jingzhong Zhang - Solving Problems in Point Geometry (2024)
100% (1)
Jingzhong Zhang - Solving Problems in Point Geometry (2024)
235 pages
The Art of Inequality - Vol2
No ratings yet
The Art of Inequality - Vol2
8 pages
102 Combinatorial Problems by Titu Andreescu & Zuming Feng
No ratings yet
102 Combinatorial Problems by Titu Andreescu & Zuming Feng
148 pages
The Art of Inequality Vol 1
No ratings yet
The Art of Inequality Vol 1
5 pages
101 Problems
No ratings yet
101 Problems
156 pages
Inequalities Marathon
No ratings yet
Inequalities Marathon
55 pages
Inequalities Marathon
No ratings yet
Inequalities Marathon
55 pages
Winter School
100% (1)
Winter School
51 pages
PVTmathscope 1376530257
No ratings yet
PVTmathscope 1376530257
56 pages
Imo 2010 - 2016
100% (1)
Imo 2010 - 2016
463 pages
Grammar and Language Workbook, Grade 9 - Glencoe - PDF Room
No ratings yet
Grammar and Language Workbook, Grade 9 - Glencoe - PDF Room
4 pages
BỘ ĐỀ THI CUỐI HK2 LỚP 9 FORM 2025 ANH 9 CUỐI KÌ 2- ỨNG HOÈ.
No ratings yet
BỘ ĐỀ THI CUỐI HK2 LỚP 9 FORM 2025 ANH 9 CUỐI KÌ 2- ỨNG HOÈ.
4 pages
Water and Sediment in Middle Distillate Fuels by Centrifuge: Standard Test Method For
No ratings yet
Water and Sediment in Middle Distillate Fuels by Centrifuge: Standard Test Method For
3 pages
Music and Emotions in The Brain Familiarity Matters
100% (1)
Music and Emotions in The Brain Familiarity Matters
9 pages
Batteries & Energy Storage
No ratings yet
Batteries & Energy Storage
17 pages
Islamic Schools, Movements and Sects
50% (2)
Islamic Schools, Movements and Sects
37 pages
(En) BQMS VendorPortal UserGuide - v1
No ratings yet
(En) BQMS VendorPortal UserGuide - v1
25 pages
Solar Progress, February 2010 Australian and New Zealnad Solar Energy Society
No ratings yet
Solar Progress, February 2010 Australian and New Zealnad Solar Energy Society
24 pages
Least Restrictive Environment: Midterm
No ratings yet
Least Restrictive Environment: Midterm
14 pages
Unit 3 Connect
No ratings yet
Unit 3 Connect
6 pages
Report On Nykaa's Business Model Sustainble Practises
No ratings yet
Report On Nykaa's Business Model Sustainble Practises
12 pages
Analysys Mason Research Report - FEMIP Broadband Potential 2010
No ratings yet
Analysys Mason Research Report - FEMIP Broadband Potential 2010
94 pages
Data Siswa
No ratings yet
Data Siswa
4 pages
Instant Download Think Sociology 2nd Edition Carl Test Bank PDF Full Chapter
100% (24)
Instant Download Think Sociology 2nd Edition Carl Test Bank PDF Full Chapter
32 pages
(English) Online Deutsch Lernen - A1.1 Anfänger - Sich Vorstellen - Das Alphabet - Die Zahlen - Aussprache (DownSub - Com)
No ratings yet
(English) Online Deutsch Lernen - A1.1 Anfänger - Sich Vorstellen - Das Alphabet - Die Zahlen - Aussprache (DownSub - Com)
8 pages
15Cs314J - Compiler Design: Unit Iii
No ratings yet
15Cs314J - Compiler Design: Unit Iii
69 pages
Rexroth Directional Valve 4we6j62
No ratings yet
Rexroth Directional Valve 4we6j62
12 pages
(Ebook PDF) Focus On Grammar 4 With Essential Online Resources 5Th Edition Download
No ratings yet
(Ebook PDF) Focus On Grammar 4 With Essential Online Resources 5Th Edition Download
56 pages
(Magic - The Gathering, #2) Wizards of The Coast, Jay Annelli - Magic - The Gathering - Legends - A Visual History-Abrams ComicArts (2020)
No ratings yet
(Magic - The Gathering, #2) Wizards of The Coast, Jay Annelli - Magic - The Gathering - Legends - A Visual History-Abrams ComicArts (2020)
268 pages
4 - Dimensional Conjunto - UL-FM (ZW10x6x12F e JU6H - UF94)
No ratings yet
4 - Dimensional Conjunto - UL-FM (ZW10x6x12F e JU6H - UF94)
2 pages
Transparent Crystal Film Screen P6 Quotation
No ratings yet
Transparent Crystal Film Screen P6 Quotation
1 page
01B Balun - Reflections 2nd Edition Chapter 21 - Some Aspects of The Balun Problem (By Walter Maxwell W2DU)
100% (1)
01B Balun - Reflections 2nd Edition Chapter 21 - Some Aspects of The Balun Problem (By Walter Maxwell W2DU)
15 pages
Rubber Dam - 1
No ratings yet
Rubber Dam - 1
69 pages
PFSR 15 Ast Pfrs 15
No ratings yet
PFSR 15 Ast Pfrs 15
30 pages
Deixis Anaphora Cataphora
No ratings yet
Deixis Anaphora Cataphora
2 pages
(Apsari Artisan) Menu Online Drink
No ratings yet
(Apsari Artisan) Menu Online Drink
9 pages
Relative Clauses - Important
No ratings yet
Relative Clauses - Important
4 pages
02 Probability of Compound Events
No ratings yet
02 Probability of Compound Events
10 pages