Microeconomics II

Spring 2021
PROF. SWAGATA BHATTACHARJEE
Today
 Normal form representation of a Game
 Examples
 Extensive form representation
 How to switch back and forth from Normal to extensive form
Representation of Games
Normal Form Games
 In the normal form representation, a game is a tuple:

 Where I= set of players i=1,2,..,n

 For each player i, the set of pure strategies (actions) available is: a particular strategy from this
set is
 Then, is the set of strategy profiles for all the players
 For each player i, the payoff function is
 Payoffs are interdependent. If player i takes strategy and other players take strategy , the payoff
to player i is:

 Then, is the set of all payoffs
Normal Form Representation of
Example 3: Prisoners’ Dilemma
Adam\Bob Deny Confess
Deny -1,-1 -3,0
Confess 0,-3 -2,-2

 Take an element . For i=A,

 Game:
 If S is finite, we can use a matrix to represent the game G
Example
 2 players (1 and 2) playing Rock, Paper, Scissors
 Rock wins over scissors, scissors wins over paper, paper wins over rock, otherwise draw.
 Winner gets 1, draw gets 0, loser gets -1.
 I=?
Example
 2 players (1 and 2) playing Rock, Paper, Scissors
 Rock wins over scissors, scissors wins over paper, paper wins over rock, otherwise draw.
 Winner gets 1, draw gets 0, loser gets -1.
 I = {1,2}

 S= ?
Example
 2 players (1 and 2) playing Rock, Paper, Scissors
 Rock wins over scissors, scissors wins over paper, paper wins over rock, otherwise draw.
 Winner gets 1, draw gets 0, loser gets -1.
 I = {1,2}

 S={RR, RP, RC, PR, PP, PC, CR, CP, CC}

1\2 R P C
R 0,0 -1,1 1,-1
P 1,-1 0,0 -1,1
C -1,1 1,-1 0,0
Assumptions
 Common Knowledge: For each player i, G is known.
 Rationality: Each player i tries to maximize ui given what he/she thinks the others will do.
 Common Knowledge of Rationality: Each player i knows that other players are rational, and that
the other players know that i is rational, and that i knows they are rational, and so on…
Review: Taxonomy of Games
 Complete and Perfect Information:
Adam decides whether to confess or deny, Bob observes what Adam has done, Bob chooses whether
to confess or deny. The game and all the payoffs are common knowledge.
 Complete and Imperfect Information:
Adam and Bob decide simultaneously whether to confess or deny. The game and all the payoffs are
common knowledge.
 Incomplete and Perfect Information:
Adam decides whether to confess or deny, Bob observes what Adam has done, but Bob does not know
what Adam’s payoffs are (for different types of people the punishment/ the act of confession mean
differently).
 Incomplete and Imperfect Information:
Adam and Bob simultaneously decide whether to confess or deny, but Bob does not know what Adam’s
payoffs are (for different types of people the punishment/ the act of confession mean differently).
Exercise
 Represent example 1 in the normal form.
Some Classic Games:
Matching Pennies
 Two players simultaneously and independently select “heads” or “tails” by each uncovering a
penny in his hand.
 If their selections match, then player 2 must give his penny to player 1; otherwise, player 1 gives
his penny to player 2.
 Zero sum games
Some Classic Games:
Matching Pennies
 Two players simultaneously and independently select “heads” or “tails” by each uncovering a
penny in his hand.
 If their selections match, then player 2 must give his penny to player 1; otherwise, player 1 gives
his penny to player 2.

I={1,2}

S={HH,HT,TH,TT}

u(HH)=(1,-1)
Coordination Game
 Both players obtain a positive payoff if they select the same strategy; otherwise they get nothing.
Stag Hunt Game
 “Pareto coordination” game or Stag Hunt Game: Two players go hunting. They can try to hunt a
stag (strategy A), or they can look for a hare (strategy B).
 If both try to hunt a stag (both choose strategy A), only then they can hunt it, in which case they
get larger share of meat. If one person looks for a stag and another looks for a hare, then they can
catch neither, and get 0. If both hunt a hare (both strategy B) then they can catch it but get a
lower share of meat.
Battle of Sexes
 Two friends have to decide whether to see a movie or go to the opera.
 Unfortunately, they work in different parts of the city and, owing to a massive outage of the
telephone system, find themselves incommunicado.
 They must simultaneously and independently select an event to attend.
 There is only one movie theater and only one opera venue, so the friends will meet each other if
they manage to coordinate their decisions.
 Both prefer to be together, regardless of which event they attend.
 However, player 1 prefers the opera and player 2 prefers the movie
Battle of Sexes
 Two friends simultaneously and independently select an event to attend.
 Both prefer to be together, regardless of which event
they attend.
 However, player 1 prefers the opera and player 2 prefers
the movie
Hawk And Dove/ Chicken
 Two players drive cars toward each other at top speed.
 Just before they reach each other, each chooses between maintaining course (H) and swerving
(D).
 If both swerve, they both save face and are satisfied.
 If only one swerves, then he is proved to be a wimp, whereas the other is lauded as a tough guy
with steely nerves.
 If both maintain course, they crash and are severely injured.
Extensive Form Representation
 Game Tree: A tree is a set of nodes connected with directed arcs such that
 There is an initial node;
 For each other node, there is one incoming arc;
 Each node can be reached through a unique path.
 Terminal nodes: no outgoing arc
A Tree
Extensive Form Game: Definition
 A game consists of
 A set of players
 A tree (to depict the sequence of moves)
 An allocation of each non-terminal node to a player (each node is a decision point for at least one
player)
 An informational partition (to be made precise)
 A payoff for each player at each terminal node.
 Extensive Form Game: Example
 Suppose there are 2 players. Player 1 goes first and chooses between Black (B) or White (W). Player 2
observes what 1 has chosen. If Player 1 has chosen B, 2 can choose between Red (R) or Yellow (Y). If 1
chose W, 2 can choose between Purple (P), Orange (O), or Green (G).
 Payoffs: if 1 chooses B, 2 chooses R, 1 gets 5, 2 gets 0
if 1 chooses B, 2 chooses Y, 1 gets 7, 2 gets 3
if 1 chooses W, 2 chooses P, 1 gets 3, 2 gets 2
if 1 chooses W, 2 chooses O, 1 gets 8, 2 gets 4
if 1 chooses W, 2 chooses G, 1 gets 0, 2 gets 7

 This situation can be depicted as an extensive form game.

I= {1,2}
 Game Tree:
Game Tree
Informational Partition
 An information set is a collection of nodes such that
 The same player is to move at each of these nodes;
 The same moves are available at each of these nodes

 Basically, the nodes in an information set are indistinguishable from the active player’s
perspective.
 An informational partition is an allocation of each non-terminal node of the tree to an
information set.
Normal Form Representation of
Example 3: Prisoners’ Dilemma
Adam\Bob Deny Confess
Deny -1,-1 -3,0
Confess 0,-3 -2,-2

 Game:
 If S is finite, we can use a matrix to represent the game G
Example 3: Prisoners’ Dilemma
 Let us represent the PD game in extensive form.
 Players: A and B
 Informational partition: because A and B move simultaneously, we can think of A moving first and then
B, but when B moves he does not know what A did.
 So, the game tree may look like:
  Game Tree:
 Nodes: 7 nodes
 Terminal nodes: 4, 5, 6,
7
Prisoners’ Dilemma  Initial node: 1
 Conventionally, nodes
are labelled by the
player active on that
node. Eg: Node 1 is
labelled A, node 2 and 3
labelled B
 Payoffs: first entry for
A, second one for B.
  Game Tree:
 Information set: Node 2
and 3 form a single
information set. Same
Prisoners’ Dilemma player (B) plays in both
nodes, same moves (C
and D) available to him,
he does not know
whether A has
Confessed or Denied.
 Typically nodes in the
same information set
are joined by dots (as in
here), or by a circle
(like the red one next
slide). (Either of them
is fine.)
  Game Tree:
 Information set: Node 2
and 3 form a single
information set. Same
Prisoners’ Dilemma player (B) plays in both
nodes, same moves (C
and D) available to him,
he does not know
whether A has
Confessed or Denied.
 Information partitions
of this tree:
{{1}, {2,3}, {4}, {5}, {6},
{7}}
Example: Senate Race
 An incumbent senator (from a rightist party) runs against a challenger (from a leftist party). The senator
has to move first and choose a political platform, leftist or rightist. If both choose the same platform, the
incumbent wins, otherwise the challenger wins. Assume that the value of winning is 10, the value of
compromising their political views (by choosing a platform not consistent with them) is 5, and the payoff
is the sum of these values.
Example: Simultaneous Legislators’ Vote
Two legislators vote whether they allow themselves a raise in salary of $2000 per year.
 Only if both vote for the salary hike, the hike proposal is accepted.
 Since voters are observing the vote, there is some loss of face for a legislator to vote for a raise.
 Let’s assume that the legislators estimate that loss of face is worth $1000 per year.
 What’s the normal form for this?
Example: Sequential Legislators’ Vote
 Three legislators vote whether they allow themselves a salary raise of $2000 per year. Since voters are
observing the vote, a legislator would estimate the loss of face by having to vote for a raise as $1000 per
year.
 If 2 out of 3 legislators vote for hike, it’s accepted, otherwise there is no hike.
 A has to vote first, then B, then C, and all votes are open.
 Rewrite the payoffs as: if there is a hike and you voted for hike, payoff is 2, if there is a hike but you did
not vote for it, payoff is 3. If there is no hike and you voted for hike, payoff is 0, if there is no hike and
you did not vote for hike, payoff is 1. (This is called normalization of payoffs)
Example: Sequential Legislators’ Vote
Example: Sequential Legislators’ Vote
 Three legislators vote whether they allow themselves a salary raise of $2000 per year. Since voters are
observing the vote, a legislator would estimate the loss of face by having to vote for a raise as $1000 per
year.
 If 2 out of 3 legislators vote for hike, it’s accepted, otherwise there is no hike.
 A has to vote first, then B, then C, and all votes are open.
 Rewrite the payoffs as: if there is a hike and you voted for hike, payoff is 2, if there is a hike but you did
not vote for it, payoff is 3. If there is no hike and you voted for hike, payoff is 0, if there is no hike and
you did not vote for hike, payoff is 1. (This is called normalization of payoffs)
Example: Sequential Legislators’ Vote
Example: Sequential Legislators’ Vote
 I={A,B,C}
Example: Sequential Legislators’ Vote
 I={A, B, C}

 WRONG!
 From B’s point of view, and are very
different, so taking action R at is NOT the
same as taking action R at !
 Remember , strategy is a complete
contingent action plan. Strategy is NOT
synonymous to Action!
Example: Sequential Legislators’ Vote
 I={A,B, C}

 Alternative way: at name the actions R and

NR, at name them R’, NR’.
 Then,
}
Example: Sequential Legislators’ Vote
 I={A,B,C}

 Alternative way: at name the actions R and

NR, at name them R’, NR’.
Example: Pricing Game Variations
 Two firms compete by each selecting either a high (H) or low (L) price for a product that they both
produce.
 Consider the following variants: Variant 1: Firm 1 moves first, Firm 2 observes and then moves.
 Notice that actions for 2
have been labelled
differently to remind that
choosing H in the top
information set is different
than choosing H in the
bottom information set.
Example: Pricing Game Variations
Normal Form

1\2 HH’ LH’ HL’ LL’

(H if 1 chooses H (L if 1 chooses H (H if 1 chooses H (L if 1 chooses H
H if 1 chooses L) H if 1 chooses L) L if 1 chooses L) L if 1 chooses L)

H 1,1 0,2 1,1 0,2

L 2,0 2,0 ½, ½ ½, ½
Example: Pricing Game Variations
 Two firms compete by each selecting either a high (H) or low (L) price for a product that they both
produce.
 Consider the following variants: Variant 2: firms move simultaneously.
Exercise
 A game is given in the extensive form. Is the following normal form representation equivalent?
If not, point out the reason(s).
Extensive Vs Normal Form: Answer
Example: KE Game
(W: Figure 2.6 and Guided Exercise Ch 3)
Example: KE Game
(W: Figure 2.6 and Guided Exercise Ch 3)
 I={K,E}
Example: KE Game
(W: Figure 2.6 and Guided Exercise Ch 3)
 I={K,E}

 Remember , strategy is a complete contingent

action plan. Strategy is NOT synonymous to
Action!
 Have to specify the entire plan for K, whenever
he gets to decide.
 That means, even if he chooses S and effectively
ends the game (so, action= S), to specify his
“strategy” we have to specify what he would be
choosing if he were to choose in the other two
nodes as well.
 Similarly, if he decides N in the second decision
point, we still have to specify what he would
choose should he produced the film.
Example: KE Game
(W: Figure 2.6 and Guided Exercise Ch 3)
Example: KE Game
(W: Figure 2.6 and Guided Exercise Ch 3)
Example: Pricing Game Variations
 Two firms compete by each selecting either a high (H) or low (L) price for a product that they both
produce.
 Consider the following variants: Variant 1: Firm 1 moves first, Firm 2 observes and then moves.
 Notice that actions for 2
have been labelled
differently to remind that
choosing H in the top
information set is different
than choosing H in the
bottom information set.
Example: Pricing Game Variations
Normal Form

1\2 HH’ LH’ HL’ LL’

(H if 1 chooses H (L if 1 chooses H (H if 1 chooses H (L if 1 chooses H
H if 1 chooses L) H if 1 chooses L) L if 1 chooses L) L if 1 chooses L)

H 1,1 0,2 1,1 0,2

L 2,0 2,0 ½, ½ ½, ½
Example: Pricing Game Variations
 Two firms compete by each selecting either a high (H) or low (L) price for a product that they both
produce.
 Consider the following variants: Variant 2: firms move simultaneously.
Next Class
 Mixed Strategy and Expected payoffs