Dynamic Games of

Complete Information

Dynamic Games of Complete and Perfect

Information

Game Theory – Fabrice Valognes -- Chp 15 1

Outline of dynamic games of complete
information
  Dynamic games of complete information
  Extensive-form representation
  Dynamic games of complete and perfect information
  Game tree
  Subgame-perfect Nash equilibrium
  Backward induction
  Applications
  Dynamic games of complete and imperfect
information
  More applications
  Repeated games

Game Theory – Fabrice Valognes -- Chp 15 2

Today’s Agenda
  Review of previous class
  Subgame-perfect Nash equilibrium
  Backward induction
  Stackelberg’s model of duopoly (2.1.B of
Gibbons, 6.2 of Osborne)
  Sequential-move Bertrand model of duopoly
(differentiated products)

Game Theory – Fabrice Valognes -- Chp 15 3

Dynamic games of complete and
perfect information
  Perfect information
  Allprevious moves are observed before
the next move is chosen.
  A player knows Who has moved What
before she makes a decision

Game Theory – Fabrice Valognes -- Chp 15 4

 Strategy and payoff
a strategy for Player 1
  A strategy for a player
player 1: H
is a complete plan of H T
actions.
  It specifies a feasible Player 2 Player 2
action for the player in
H T H T
every contingency in
which the player might
be called on to act. -1, 1 1, -1 1, -1 -1, 1
  It specifies what the
a strategy for player 2: H if player 1 plays
player does at each of H, T if player 1 plays T (written as HT)
her nodes
Player 1’s payoff is -1 and player 2’s payoff is
1 if player 1 plays H and player 2 plays HT
Game Theory – Fabrice Valognes -- Chp 15 5
Nash equilibrium in a dynamic game
  We can also use normal-form to represent a
dynamic game
  The set of Nash equilibria in a dynamic game
of complete information is the set of Nash
equilibria of its normal-form
  How to find the Nash equilibria in a dynamic
game of complete information
  Construct the normal-form of the dynamic
game of complete information
  Find the Nash equilibria in the normal-form

Game Theory – Fabrice Valognes -- Chp 15 6

Subgame-perfect Nash equilibrium
  A Nash equilibrium of a dynamic game is
subgame-perfect if the strategies of the Nash
equilibrium constitute or induce a Nash
equilibrium in every subgame of the game.

  Subgame-perfect Nash equilibrium is a Nash

equilibrium.

Game Theory – Fabrice Valognes -- Chp 15 7

 Subgame
  A subgame of a game tree Player 1
begins at a nonterminal node
and includes all the nodes and H T
edges following the
nonterminal node Player 2 Player 2
  A subgame beginning at a
nonterminal node x can be H T H T
obtained as follows:
  remove the edge -1, 1 1, -1 1, -1 -1, 1
connecting x and its
predecessor
  the connected part
containing x is the subgame a subgame

Game Theory – Fabrice Valognes -- Chp 15 8

Existence of subgame-perfect Nash
equilibrium
  Every finite dynamic game of complete and
perfect information has a subgame-perfect
Nash equilibrium that can be found by
backward induction.

Game Theory – Fabrice Valognes -- Chp 15 9

Backward induction: illustration

Player 1
C D

Player 2 Player 2
E F G H

2, 1 3, 0 0, 2 1, 3

  Subgame-perfect Nash equilibrium (C, EH).

  player 1 plays C;
  player 2 plays E if player 1 plays C, plays H if player 1
plays D.
Game Theory – Fabrice Valognes -- Chp 15 10
 Multiple subgame-perfect Nash
equilibria: illustration
Player 1
C E
D

Player 2 Player 2 Player 2

F G H I J K

0, 1 1, 0 1, 1 2, 1 2, 2 1, 3

  Subgame-perfect Nash equilibrium (D, FHK).

  player 1 plays D
  player 2 plays F if player 1 plays C, plays H if player 1
plays D, plays K if player 1 plays E.
Game Theory – Fabrice Valognes -- Chp 15 11
 Multiple subgame-perfect Nash
equilibria
Player 1
C E
D

Player 2 Player 2 Player 2

F G H I J K

0, 1 1, 0 1, 1 2, 1 2, 2 1, 3

  Subgame-perfect Nash equilibrium (E, FHK).

  player 1 plays E;
  player 2 plays F if player 1 plays C, plays H if player 1
plays D, plays K if player 1 plays E.
Game Theory – Fabrice Valognes -- Chp 15 12
 Multiple subgame-perfect Nash
equilibria
Player 1
C E
D

Player 2 Player 2 Player 2

F G H I J K

0, 1 1, 0 1, 1 2, 1 2, 2 1, 3

  Subgame-perfect Nash equilibrium (D, FIK).

  player 1 plays D;
  player 2 plays F if player 1 plays C, plays I if player 1
plays D, plays K if player 1 plays E.
Game Theory – Fabrice Valognes -- Chp 15 13
Stackelberg model of duopoly
  A homogeneous product is produced by only two
firms: firm 1 and firm 2. The quantities are denoted by
q1 and q2, respectively.
  The timing of this game is as follows:
  Firm 1 chooses a quantity q1 ≥0.
  Firm 2 observes q1 and then chooses a quantity q2 ≥0.
  The market priced is P(Q)=a –Q, where a is a
constant number and Q=q1+q2.
  The cost to firm i of producing quantity qi is Ci(qi)=cqi.
  Payoff functions:
u1(q1, q2)=q1(a–(q1+q2)–c)
u2(q1, q2)=q2(a–(q1+q2)–c)
Game Theory – Fabrice Valognes -- Chp 15 14
Stackelberg model of duopoly
  Find the subgame-perfect Nash equilibrium
by backward induction
  We first solve firm 2’s problem for any q1 ≥0 to
get firm 2’s best response to q1 . That is, we
first solve all the subgames beginning at firm
2.
  Then we solve firm 1’s problem. That is, solve
the subgame beginning at firm 1

Game Theory – Fabrice Valognes -- Chp 15 15

Stackelberg model of duopoly
  Solve firm 2’s problem for any q1 ≥0 to get firm 2’s
best response to q1.
  Max u2(q1, q2)=q2(a–(q1+q2)–c)
subject to 0 ≤ q2 ≤ +∞

FOC: a – 2q2 – q1 – c = 0

  Firm 2’s best response,

  R2(q1) = (a – q1 – c)/2 if q1 ≤ a– c
= 0 if q1 > a– c

Note: Osborne used b2(q1) instead of R2(q1)

Game Theory – Fabrice Valognes -- Chp 15 16
Stackelberg model of duopoly
  Solve firm 1’s problem. Note firm 1 can also solve
firm 2’s problem. That is, firm 1 knows firm 2’s best
response to any q1. Hence, firm 1’s problem is
  Max u1(q1, R2(q1))=q1(a–(q1+R2(q1))–c)
subject to 0 ≤ q1 ≤ +∞

That is,
Max u1(q1, R2(q1))=q1(a–q1–c)/2
subject to 0 ≤ q1 ≤ +∞

FOC: (a – 2q1 – c)/2 = 0

q1 = (a – c)/2

Game Theory – Fabrice Valognes -- Chp 15 17

Stackelberg model of duopoly
  Subgame-perfect Nash equilibrium
  ( (a – c)/2, R2(q1) ), where
R2(q1) = (a – q1 – c)/2 if q1 ≤ a– c
= 0 if q1 > a– c
  That is, firm 1 chooses a quantity (a – c)/2, firm
2 chooses a quantity R2(q1) if firm 1 chooses a
quantity q1.
  The backward induction outcome is ( (a – c)/2,
(a – c)/4 ).
  Firm 1 chooses a quantity (a – c)/2, firm 2
chooses a quantity (a – c)/4.
Game Theory – Fabrice Valognes -- Chp 15 18
Stackelberg model of duopoly
  Firm 1 produces
q1=(a – c)/2 and its profit
q1(a–(q1+ q2)–c)=(a–c)2/8

  Firm 2 produces
q2=(a – c)/4 and its profit
q2(a–(q1+ q2)–c)=(a–c)2/16

  The aggregate quantity is 3(a – c)/4.

Game Theory – Fabrice Valognes -- Chp 15 19

Cournot model of duopoly
  Firm 1 produces
q1=(a – c)/3 and its profit
q1(a–(q1+ q2)–c)=(a–c)2/9

  Firm 2 produces
q2=(a – c)/3 and its profit
q2(a–(q1+ q2)–c)=(a–c)2/9

  The aggregate quantity is 2(a – c)/3.

Game Theory – Fabrice Valognes -- Chp 15 20

Monopoly
  Suppose that only one firm, a monopoly, produces
the product. The monopoly solves the following
problem to determine the quantity qm.
  Max qm (a–qm–c)
subject to 0 ≤ qm ≤ +∞

FOC: a – 2qm – c = 0
qm = (a – c)/2
  Monopoly produces
qm=(a – c)/2 and its profit
qm(a–qm–c)=(a–c)2/4

Game Theory – Fabrice Valognes -- Chp 15 21

Sequential-move Bertrand model of
duopoly (differentiated products)
  Two firms: firm 1 and firm 2.
  Each firm chooses the price for its product. The
prices are denoted by p1 and p2, respectively.
  The timing of this game as follows.
  Firm 1 chooses a price p1 ≥0.
  Firm 2 observes p1 and then chooses a price p2 ≥0.
  The quantity that consumers demand from firm 1:
q1(p1, p2) = a – p1 + bp2.
  The quantity that consumers demand from firm 2:
q2(p1, p2) = a – p2 + bp1.
  The cost to firm i of producing quantity qi is Ci(qi)=cqi.
Game Theory – Fabrice Valognes -- Chp 15 22
Sequential-move Bertrand model of
duopoly (differentiated products)
  Solve firm 2’s problem for any p1 ≥0 to get
firm 2’s best response to p1.
  Max u2(p1, p2)=(a – p2 + bp1 )(p2 – c)
subject to 0 ≤ p2 ≤ +∞

FOC: a + c – 2p2 + bp1 = 0

p2 = (a + c + bp1)/2
  Firm 2’s best response,
  R2(p1) = (a + c + bp1)/2

Game Theory – Fabrice Valognes -- Chp 15 23

Sequential-move Bertrand model of
duopoly (differentiated products)
  Solve firm 1’s problem. Note firm 1 can also solve
firm 2’s problem. Firm 1 knows firm 2’s best response
to p1. Hence, firm 1’s problem is
  Max u1(p1, R2(p1))=(a – p1 + b×R2(p1) )(p1 – c)
subject to 0 ≤ p1 ≤ +∞

That is,
Max u1(p1, R2(p1))=(a – p1 + b×(a + c + bp1)/2 )(p1 – c)
subject to 0 ≤ p1 ≤ +∞

  FOC: a – p1 + b×(a + c + bp1)/2+(–1+b2/2) (p1 – c) = 0

p1 = (a+c+(ab+bc–b2c)/2)/(2–b2)

Game Theory – Fabrice Valognes -- Chp 15 24

Sequential-move Bertrand model of duopoly
(differentiated products)
  Subgame-perfect Nash equilibrium
2
  ((a+c+(ab+bc–bc)/2)/(2–b2), R2(p1) ),
where R2(p1) = (a + c + bp1)/2

  Firm 1 chooses a price

(a+c+(ab+bc–b2c)/2)/(2–b2),
firm 2 chooses a price R2(p1) if firm 1 chooses
a price p1.

Game Theory – Fabrice Valognes -- Chp 15 25

Summary
  Subgame perfect Nash equilibrium
  Backward induction
  Stackelberg model of duopoly
  Sequential-move Bertrand model of duopoly
(differentiated products)

  Next time
  Dynamic games of complete and imperfect information

  Reading lists
  Sec 2.2, 2.4 of Gibbons

Game Theory – Fabrice Valognes -- Chp 15 26