Proofs

Induction
In previous notes you learned how to prove results by direct proof, indirect proof, proof by
contradiction, or proof by cases. In these notes you will learn another method of proof, induction.
Induction is used to show that a large summation is always true.

1 Summation notation
Suppose I want to add the numbers from 1 to 1000. If I try to write this out properly then I’ll have
to write 1 + 2+ with 1000 terms.
1 + 2 + 3 + 4 + 5 + 6 + 7 + 8 + yuck!
I can’t write 1000 numbers! A more concise way of writing this is to use “summation notation”
with the uppercase Greek letter Sigma Σ. For example,
1000
X
k
k=1

This means that you add the term k repeatedly starting with k = 1 until you reach k = 1000.
Another example is
X2
(3k + 1)
k=0
This means that you add the expression (3k + 1) repeatedly starting with k = 0 until you reach
k = 2. In particular,
2
X
(3k + 1) = (3 · 0 + 1) + (3 · 1 + 1) + (3 · 2 + 1) = 12
k=0

If you like writing sums this way then you are welcome to do so. However, I think that most
students would prefer to avoid this extra abstraction. Instead I will use the + · · · + notation to
indicate that a sum is taking place.
Important – Summation Notation: The + · · · + notation means that that you add up numbers
following the given pattern until you reach the final term in the sequence. For example:
• Sum 1 + 4 + 7 + · · · + (3n + 1) with n = 2.
Add until you reach (3n + 1) = (3 ∗ 2 + 1) = 7, so
if n = 2 then 1 + 4 + 7 + · · · + (3n + 1) = 1 + 4 + 7 = 12

• Sum 1 + 2 + 3 + · · · + n with n = 2.
Add until you reach (n) = 2, so
if n = 2 then 1 + 2 + 3 + · · · + n = 1 + 2 = 3

1
2 Induction
Legend has it that when the genius mathematician Gauss was a child his teacher tried to shut
him up by telling him to add the numbers from 1 to 1000. A minute later Gauss gave the right
answer. The same legend says that the teacher eventually had a nervous breakdown and was
institutionalized because he couldn’t cope with a kid genius being smarter than him.
Did Gauss add up 1000 numbers in only a minute? This is very unlikely. Instead, he probably
figured out the following formula:

n(n + 1)
if n ≥ 1 then (1 + 2 + 3 + · · · + n) = (1)
2
n(n+1) 1000·1001
Gauss wanted to add up to n = 1000, so his answer would be 2 = 2 = 500500.
Here are a few more formulas for summations:

• if n ≥ 1 then 1 + 3 + 5 + · · · + (2n − 1) = n2
This says that the sum of the first n odd numbers is n2 .
1 1 1 1 1
• if n ≥ 0 then 1+ 2 + 4 + 8 + ··· + 2n =2− 2n

• if n ≥ 0 then 1 + 2 + 4 + 8 + · · · + 2n = 2n+1 − 1

How does one show that a sum like this is always true? It’s not good enough to just check a
few numbers and declare that the formula seems to work. Instead, we use a method of proof that
is related to knocking down dominoes:

• First, push the first domino.

• Second, observe as each domino knocks down the next one.

• End result: All dominoes will fall down.

The same idea can be applied to proving that a mathematical claim is true for all values of n.

• First, show that the claim is true for the smallest possible value of n.

• Second, show that each time the claim is true then it causes the next case to be true too.

• End result: The claim is true for all values of n.

More formally:

2
Definition 2.1. An Inductive proof has two parts.

1. Base case: Show that the result holds for the simplest case. This is often n = 0 or n = 1, but
not always. This is like “pushing” down the first domino. For Gauss’s result this is

W hen n = 1 then 1 + 2 + 3 + · · · + n = 1
n(n + 1) 1 · (1 + 1)
= = 1
2 2
T hey 0 re equal!

2. Inductive step: Assume that the result holds when n = k. Write out exactly what this means,
such as
k(k + 1)
Assume : 1 + 2 + 3 + · · · + k = (2)
2
Use the n = k case to show the n = k + 1 case is also true. Write the (k + 1) case precisely.
Always list the last two terms:

(k + 1)(k + 2)
Show : 1 + 2 + 3 + · · · + k + (k + 1) = (3)
2

3. To prove that (3) is true, first observe that it starts with the same equation that is in (2).
Substitute the formula from (2), and do algebra to simplify and get the term on the right.

1 + 2 + 3 + · · · + k + (k + 1) = |1 + 2 + 3{z+ · · · + k} +(k + 1)
equation (2)
k(k + 1)
= + (k + 1)
2  
k (k + 1)(k + 2)
= (k + 1) +1 =
2 2

4. If this were dominoes then we could conclude that every domino will fall. Since this is math
we can instead conclude that the formula we wanted is true for all values.
n(n + 1)
Conclusion : ∀n ≥ 1, (1 + 2 + 3 + · · · + n) =
2

3
 We will do a few examples. I will start by repeating what was found on the previous page, but
this time I present only the portions relevant to the proof.

Important: You must include all of these details in your proofs. If you just write down formulas
and leave out the explanation then you don’t have a proof, you just have a collection of formulas
on a page:

Example 2.2. Show that

n(n + 1)
∀n ≥ 1, (1 + 2 + 3 + · · · + n) =
2
Proof. Base Case: The smallest case is n = 1. In this case the claim is that

1(1 + 1)
1=
2
Indeed, 1 = 1.

Inductive Step: Assume the result holds when n = k, i.e.

k(k + 1)
1 + 2 + 3 + ··· + k = (4)
2
Show that the result holds when n = k + 1, i.e.

(k + 1)(k + 2)
1 + 2 + 3 + · · · + k + (k + 1) = (5)
2
Starting with the left hand expression we substitute in the assumed case (“inductive hypothe-
sis”) and after some algebra we should arrive at the desired result.

k(k + 1)
|1 + 2 + 3{z+ · · · + k} +(k + 1) = 2
+ (k + 1) inductive hypothesis (4)
k+1
= (k + 2) factor out largest term, and denominator
2
The final equality is exactly the desired result (equation (5)), and so Gauss’s result is proven.

At the level of this course inductive proofs almost always follow the same format.

4
Here is another example:

Example 2.3. An n-bit binary string is a sequence of n 0’s and 1’s. For example, 011001 is a 6
bit binary string, and the empty string “” is a 0 bit binary string.
Q: How many distinct binary strings use ≤ n bits, for n ≥ 0?
A: The number of n bit strings is 2n . It follows that the number of binary strings with k ≤ n
bits is
Xn
2k = 20 + 21 + 22 + · · · + 2n
k=0

How large is this sum? Use induction to show that

20 + 21 + 22 + · · · + 2n = 2n+1 − 1

Proof. Base Case: The smallest case is n = 0. In this case the claim is that

20 = 1 = 20+1 − 1

Inductive Step: Assume the result holds when n = k, i.e.

20 + 21 + 22 + 23 + · · · + 2k = 2k+1 − 1 (6)

Show that the result holds when n = k + 1, i.e.

20 + 21 + 22 + 23 + · · · + 2k + 2k+1 = 2k+2 − 1 (7)

Starting with the left hand expression we substitute in the assumed case (“inductive hypothe-
sis”) and after some algebra we should arrive at the desired result.
 
2| 0 + 21 + 22 +
{z 2 3
+ · · · + 2 k
} +2k+1
= 2 k+1
− 1 + 2k+1 inductive hypothesis (6)

= 2 · 2k+1 − 1 simplify
k+2
= 2 −1 2 2 = 21+a
1 a

The final equality is exactly the desired result (equation (7)).

5
For the next example we use notation that will arise a lot in the next “Theme.”

Definition 2.4. “n factorial” is the product of all integers ≤ n. In other words,

n! = n(n − 1)(n − 2) · · · 3 · 2 · 1

For instance,
5! = 5 · 4 · 3 · 2 · 1 = 120 and (n + 1)! = n! (n + 1)

We will encounter factorials a lot when we study counting problems such as the number of Full
House’s in a deck of cards. For instance, the number of possible 5 card hands is
52!
47!5!
This plays a role in finding the probability that a 5 card hand is a Full House, for instance.
Now an example with factorials.

Example 2.5. Show that if n ≥ 4 then

Proof. There is a fairly straightforward proof of this. Can you figure it out? I will give an inductive
proof instead.
Base Case: The smallest case is n = 4. In this case the claim is that

4! > 24

Indeed, 24 > 16.

Inductive Step: Assume the result holds when n = k, i.e.

k! > 2k (8)

Show that the result holds when n = k + 1, i.e.

(k + 1)! > 2k+1

Starting with the left hand expression we substitute in the assumed case (“inductive hypothesis”)
and after some algebra we should arrive at the desired result.

(k + 1)! = k! (k + 1)
> 2k (k + 1) inductive hypothesis (8)
≥ 2k 5 because k ≥ 4 and so k + 1 ≥ 5
> 2k 2 because 5 > 2.

The final equality is exactly the desired result, and so the claim is proven.

There are two many variations on these for me to cover them all but I encourage you to read
several examples in the book.