Strategic Form Games

CS535: Introduction to Game Theory and Mechanism Design

Strategic Form Games: Notation

1
Strategic Form Games: Notation

• N = {1, 2, . . . , n}: Set of players

1
Strategic Form Games: Notation

• N = {1, 2, . . . , n}: Set of players

• Si : Strategy set of player i

1
Strategic Form Games: Notation

• N = {1, 2, . . . , n}: Set of players

• Si : Strategy set of player i

• si ∈ Si : A strategy of player i

1
Strategic Form Games: Notation

• N = {1, 2, . . . , n}: Set of players

• Si : Strategy set of player i

• si ∈ Si : A strategy of player i

• S = ×ni=1 Si : Set of all possible strategy profiles

1
Strategic Form Games: Notation

• N = {1, 2, . . . , n}: Set of players

• Si : Strategy set of player i

• si ∈ Si : A strategy of player i

• S = ×ni=1 Si : Set of all possible strategy profiles

• s = (s1 , s2 , . . . , sn ): A strategy profile

1
Strategic Form Games: Notation

• N = {1, 2, . . . , n}: Set of players

• Si : Strategy set of player i

• si ∈ Si : A strategy of player i

• S = ×ni=1 Si : Set of all possible strategy profiles

• s = (s1 , s2 , . . . , sn ): A strategy profile

• ui : ×ni=1 Si → R: Utility function of player i

1
Strategic Form Games: Notation

• N = {1, 2, . . . , n}: Set of players

• Si : Strategy set of player i

• si ∈ Si : A strategy of player i

• S = ×ni=1 Si : Set of all possible strategy profiles

• s = (s1 , s2 , . . . , sn ): A strategy profile

• ui : ×ni=1 Si → R: Utility function of player i

• u1 (s1 , s2 , . . . , sn ) = α: Utility of player 1 for the given strategy profile

1
Strategic Form Games: Notation

• N = {1, 2, . . . , n}: Set of players

• Si : Strategy set of player i

• si ∈ Si : A strategy of player i

• S = ×ni=1 Si : Set of all possible strategy profiles

• s = (s1 , s2 , . . . , sn ): A strategy profile

• ui : ×ni=1 Si → R: Utility function of player i

• u1 (s1 , s2 , . . . , sn ) = α: Utility of player 1 for the given strategy profile

• S−i = ×j̸=i Sj : Set of strategy profiles of players other than player i

1
Strategic Form Games: Notation

• N = {1, 2, . . . , n}: Set of players

• Si : Strategy set of player i

• si ∈ Si : A strategy of player i

• S = ×ni=1 Si : Set of all possible strategy profiles

• s = (s1 , s2 , . . . , sn ): A strategy profile

• ui : ×ni=1 Si → R: Utility function of player i

• u1 (s1 , s2 , . . . , sn ) = α: Utility of player 1 for the given strategy profile

• S−i = ×j̸=i Sj : Set of strategy profiles of players other than player i

• s−i ∈ S−i : A strategy profile of players other than player i

1
Strategic Form Games: Notation

• N = {1, 2, . . . , n}: Set of players

• Si : Strategy set of player i

• si ∈ Si : A strategy of player i

• S = ×ni=1 Si : Set of all possible strategy profiles

• s = (s1 , s2 , . . . , sn ): A strategy profile

• ui : ×ni=1 Si → R: Utility function of player i

• u1 (s1 , s2 , . . . , sn ) = α: Utility of player 1 for the given strategy profile

• S−i = ×j̸=i Sj : Set of strategy profiles of players other than player i

• s−i ∈ S−i : A strategy profile of players other than player i

• ui (si , s−i ) : Utility of player i when it plays si and the other players
play according to strategy profile s−i
1
Questions to Answer

Given a game < N, (Si )ni=1 , (ui )ni=1 >, we need to answer the following:

2
Questions to Answer

Given a game < N, (Si )ni=1 , (ui )ni=1 >, we need to answer the following:

• Which strategy should a player never play if he/she is rational and

intelligent, under the common knowledge assumption?

2
Questions to Answer

Given a game < N, (Si )ni=1 , (ui )ni=1 >, we need to answer the following:

• Which strategy should a player never play if he/she is rational and

intelligent, under the common knowledge assumption?
• Which is the best strategy for him/her?

2
Questions to Answer

Given a game < N, (Si )ni=1 , (ui )ni=1 >, we need to answer the following:

• Which strategy should a player never play if he/she is rational and

intelligent, under the common knowledge assumption?
• Which is the best strategy for him/her?
• Does the best strategy always exist?

2
Questions to Answer

Given a game < N, (Si )ni=1 , (ui )ni=1 >, we need to answer the following:

• Which strategy should a player never play if he/she is rational and

intelligent, under the common knowledge assumption?
• Which is the best strategy for him/her?
• Does the best strategy always exist?
• If such a best strategy does not exist, what will the player do?

2
Strongly Dominated Strategy

A strategy si is strongly dominated if ∃ti ∈ Si s.t. ∀s−i ∈ S−i , we have:

ui (si , s−i ) < ui (ti , s−i )

3
Strongly Dominated Strategy

A strategy si is strongly dominated if ∃ti ∈ Si s.t. ∀s−i ∈ S−i , we have:

ui (si , s−i ) < ui (ti , s−i )

A rational, intelligent player will never play a strongly dominated strategy

3
Example: Strongly Dominated Strategy

L R
T 1,0 1,2
B 0,3 0,1

4
Example: Strongly Dominated Strategy

L R
T 1,0 1,2
B 0,3 0,1

• B is a strongly dominated strategy for player 1

4
Example: Strongly Dominated Strategy

L R
T 1,0 1,2
B 0,3 0,1

• B is a strongly dominated strategy for player 1

• Is there any strongly dominated strategy for player 2?

4
Example: Strongly Dominated Strategy

L R
T 1,0 1,2
B 0,3 0,1

• B is a strongly dominated strategy for player 1

• Is there any strongly dominated strategy for player 2?
• We can eliminate B from player 1’s strategy set

4
Example: Strongly Dominated Strategy

L R
T 1,0 1,2
B 0,3 0,1

• B is a strongly dominated strategy for player 1

• Is there any strongly dominated strategy for player 2?
• We can eliminate B from player 1’s strategy set

L R
T 1,0 1,2

4
Example: Strongly Dominated Strategy

L R
T 1,0 1,2
B 0,3 0,1

• B is a strongly dominated strategy for player 1

• Is there any strongly dominated strategy for player 2?
• We can eliminate B from player 1’s strategy set

L R
T 1,0 1,2

• In the residual game, L is a strongly dominated strategy for player 2

4
Example: Strongly Dominated Strategy

L R
T 1,0 1,2
B 0,3 0,1

• B is a strongly dominated strategy for player 1

• Is there any strongly dominated strategy for player 2?
• We can eliminate B from player 1’s strategy set

L R
T 1,0 1,2

• In the residual game, L is a strongly dominated strategy for player 2

• So, the strategy profile that will be played is (T,R)

4
Strongly Dominated Strategy

A strategy si is strongly dominated if ∃ti ∈ Si s.t. ∀s−i ∈ S−i , we have:

ui (si , s−i ) < ui (ti , s−i )

A rational, intelligent player will never play a strongly dominated strategy

L R
T 10, 0,
C 0, 10,
B 9, 9,

Does there exist any strongly dominated strategy for player 1?

5
Strongly Dominated Strategy

A strategy si is strongly dominated if ∃ti ∈ Si s.t. ∀s−i ∈ S−i , we have:

ui (si , s−i ) < ui (ti , s−i )

A rational, intelligent player will never play a strongly dominated strategy

L R
T 10, 0,
C 0, 10,
B 9, 9,

Does there exist any strongly dominated strategy for player 1?

Beware: It is ∃ti ∈ Si s.t. ∀s−i ∈ S−i and not ∀s−i ∈ S−i , ∃ti ∈ Si

5
Strongly Dominant Strategy

• si∗ is a strongly dominant strategy if it strongly dominates every

other strategy si ∈ Si , si ̸= si∗ , i.e.,

ui (si∗ , s−i ) > ui (si , s−i ), ∀s−i ∈ S−i , ∀si ∈ Si \ {si∗ }

6
Strongly Dominant Strategy

• si∗ is a strongly dominant strategy if it strongly dominates every

other strategy si ∈ Si , si ̸= si∗ , i.e.,

ui (si∗ , s−i ) > ui (si , s−i ), ∀s−i ∈ S−i , ∀si ∈ Si \ {si∗ }

• In Prisoners’ Dilemma, C is the strongly dominant strategy

6
Strongly Dominant Strategy

• si∗ is a strongly dominant strategy if it strongly dominates every

other strategy si ∈ Si , si ̸= si∗ , i.e.,

ui (si∗ , s−i ) > ui (si , s−i ), ∀s−i ∈ S−i , ∀si ∈ Si \ {si∗ }

• In Prisoners’ Dilemma, C is the strongly dominant strategy

• A player will always play strongly dominant strategy (if it exists) no
matter what other players play

6
Strongly Dominant Strategy

• si∗ is a strongly dominant strategy if it strongly dominates every

other strategy si ∈ Si , si ̸= si∗ , i.e.,

ui (si∗ , s−i ) > ui (si , s−i ), ∀s−i ∈ S−i , ∀si ∈ Si \ {si∗ }

• In Prisoners’ Dilemma, C is the strongly dominant strategy

• A player will always play strongly dominant strategy (if it exists) no
matter what other players play
• If there exists a strongly dominant strategy, then other strategies will
be strongly dominated, by definition

6
Points to Note

• If a player has two strategies, and there exists a strongly dominated

strategy, the other strategy will automatically be strongly dominant

7
Points to Note

• If a player has two strategies, and there exists a strongly dominated

strategy, the other strategy will automatically be strongly dominant
• If there are more than two strategies and there exists a strongly
dominated strategy, it does not imply that a strongly dominant
strategy exists (Example?)

7
Strongly Dominant Strategy Equilibrium (SDSE)

A strategy profile (s1∗ , s2∗ , . . . , sn∗ ) is called strongly dominant strategy

equilibrium (SDSE) if si∗ is a strongly dominant strategy for player i.

8
Strongly Dominant Strategy Equilibrium (SDSE)

A strategy profile (s1∗ , s2∗ , . . . , sn∗ ) is called strongly dominant strategy

equilibrium (SDSE) if si∗ is a strongly dominant strategy for player i.
SDSE in the Prisoners’ Dilemma game:

8
Strongly Dominant Strategy Equilibrium (SDSE)

A strategy profile (s1∗ , s2∗ , . . . , sn∗ ) is called strongly dominant strategy

equilibrium (SDSE) if si∗ is a strongly dominant strategy for player i.
SDSE in the Prisoners’ Dilemma game: (C,C)

8
SDSE

• It may be the case that there does not exist a strongly dominant
strategy for any player (Example?)

9
SDSE

• It may be the case that there does not exist a strongly dominant
strategy for any player (Example?)

L R
T 1,2 2,3
B 2,2 2,0

9
SDSE

• It may be the case that there does not exist a strongly dominant
strategy for any player (Example?)

L R
T 1,2 2,3
B 2,2 2,0

• It may be the case that there exist strongly dominant strategies for
only some players, thus not having SDSE

9
SDSE

• It may be the case that there does not exist a strongly dominant
strategy for any player (Example?)

L R
T 1,2 2,3
B 2,2 2,0

• It may be the case that there exist strongly dominant strategies for
only some players, thus not having SDSE

L R
T 1,0 1,2
B 0,3 0,1
9
Example: Tragedy of the Commons

• There are n farmers

• Si = {0, 1} : Whether to keep the sheep or not
• Keeping sheep will give 1 unit of profit but will cause 5 units of loss
to the environment, which needs to be shared among all the farmers
• Is there any strongly dominant strategy for farmer i?

10
Example: Tragedy of the Commons

• There are n farmers

• Si = {0, 1} : Whether to keep the sheep or not
• Keeping sheep will give 1 unit of profit but will cause 5 units of loss
to the environment, which needs to be shared among all the farmers
• Is there any strongly dominant strategy for farmer i?

• If n > 5 then si = 1 is strongly dominant strategy

10
Example: Tragedy of the Commons

• There are n farmers

• Si = {0, 1} : Whether to keep the sheep or not
• Keeping sheep will give 1 unit of profit but will cause 5 units of loss
to the environment, which needs to be shared among all the farmers
• Is there any strongly dominant strategy for farmer i?

• If n > 5 then si = 1 is strongly dominant strategy

• If n < 5, then si = 0 is strongly dominant strategy

10
Example: Tragedy of the Commons

• There are n farmers

• Si = {0, 1} : Whether to keep the sheep or not
• Keeping sheep will give 1 unit of profit but will cause 5 units of loss
to the environment, which needs to be shared among all the farmers
• Is there any strongly dominant strategy for farmer i?

• If n > 5 then si = 1 is strongly dominant strategy

• If n < 5, then si = 0 is strongly dominant strategy
• What happens when n = 5?

10
Example: Tragedy of the Commons

• There are n farmers

• Si = {0, 1} : Whether to keep the sheep or not
• Keeping sheep will give 1 unit of profit but will cause 5 units of loss
to the environment, which needs to be shared among all the farmers
• Is there any strongly dominant strategy for farmer i?

• If n > 5 then si = 1 is strongly dominant strategy

• If n < 5, then si = 0 is strongly dominant strategy
• What happens when n = 5?

• Does there exist an SDSE?

10
Weakly Dominant Strategy Equilibrium (WDSE)

• si∗ is a weakly dominant strategy if ∀si ∈ Si :

• ∀s−i ∈ S−i : ui (si∗ , s−i ) ≥ ui (si , s−i )
• ∃t−i ∈ S−i : ui (si∗ , t−i ) > ui (si , t−i )

11
Weakly Dominant Strategy Equilibrium (WDSE)

• si∗ is a weakly dominant strategy if ∀si ∈ Si :

• ∀s−i ∈ S−i : ui (si∗ , s−i ) ≥ ui (si , s−i )
• ∃t−i ∈ S−i : ui (si∗ , t−i ) > ui (si , t−i )
• (s1∗ , s2∗ , . . . , sn∗ ) is called weakly dominant strategy equilibirium
(WDSE) if si∗ is a weakly dominant strategy for player i

11
Weakly Dominant Strategy Equilibrium (WDSE)

• si∗ is a weakly dominant strategy if ∀si ∈ Si :

• ∀s−i ∈ S−i : ui (si∗ , s−i ) ≥ ui (si , s−i )
• ∃t−i ∈ S−i : ui (si∗ , t−i ) > ui (si , t−i )
• (s1∗ , s2∗ , . . . , sn∗ ) is called weakly dominant strategy equilibirium
(WDSE) if si∗ is a weakly dominant strategy for player i

• si∗ is a very weakly dominant strategy if ∀si ∈ Si :

• ∀s−i ∈ S−i : ui (si∗ , s−i ) ≥ ui (si , s−i )

11
Weakly Dominant Strategy Equilibrium (WDSE)

• si∗ is a weakly dominant strategy if ∀si ∈ Si :

• ∀s−i ∈ S−i : ui (si∗ , s−i ) ≥ ui (si , s−i )
• ∃t−i ∈ S−i : ui (si∗ , t−i ) > ui (si , t−i )
• (s1∗ , s2∗ , . . . , sn∗ ) is called weakly dominant strategy equilibirium
(WDSE) if si∗ is a weakly dominant strategy for player i

• si∗ is a very weakly dominant strategy if ∀si ∈ Si :

• ∀s−i ∈ S−i : ui (si∗ , s−i ) ≥ ui (si , s−i )

Strongly dominant =⇒ Weakly dominant =⇒ Very weakly dominant

11
Example: WDSE

L R
T 1,2 2,3
B 2,2 2,0

12
Example: WDSE

L R
T 1,2 2,3
B 2,2 2,0

• B is a weakly dominant strategy for player 1

• No strategy is a weakly dominant strategy for player 2
• So, there does not exist any WDSE

12
Uniqueness

• If a strongly dominant strategy exists for a player, it is unique

13
Uniqueness

• If a strongly dominant strategy exists for a player, it is unique

• If a weakly dominant strategy exists for a player, it is unique

13
Order of Elimination of Dominated Strategies

• Rational players will not play strongly dominated strategies, so

eliminate them to reach the residual game
• The order in which strongly dominated strategies are eliminated
does not matter, i.e., the same residual game is obtained (Proof?)
• Iterated removal of weakly dominated strategies may result in
different residual games with different orders of elimination

14
Example: Iterated Elimination of Dominated Strategies

L C R
T 1,2 2,3 0,3
M 2,2 2,1 3,2
B 2,1 0,0 1,0

15
Example: Iterated Elimination of Dominated Strategies

L C R
T 1,2 2,3 0,3
M 2,2 2,1 3,2
B 2,1 0,0 1,0

• Order of Elimination: T, R, B, C → (M,L)

15
Example: Iterated Elimination of Dominated Strategies

L C R
T 1,2 2,3 0,3
M 2,2 2,1 3,2
B 2,1 0,0 1,0

• Order of Elimination: T, R, B, C → (M,L)

• Order of Elimination: B, L, C, T → (M,R)

15
Example: Iterated Elimination of Dominated Strategies

L C R
T 1,2 2,3 0,3
M 2,2 2,1 3,2
B 2,1 0,0 1,0

• Order of Elimination: T, R, B, C → (M,L)

• Order of Elimination: B, L, C, T → (M,R)
• Order of Elimination: T, C, R → (M,L) or (B,L)

15
Example: Iterated Elimination of Dominated Strategies

L C R
T 1,2 2,3 0,3
M 2,2 2,1 3,2
B 2,1 0,0 1,0

• Order of Elimination: T, R, B, C → (M,L)

• Order of Elimination: B, L, C, T → (M,R)
• Order of Elimination: T, C, R → (M,L) or (B,L)
• Does there exist a WDSE?

15
Example: Iterated Elimination of Dominated Strategies

L C R
T 1,2 2,3 0,3
M 2,2 2,1 3,2
B 2,1 0,0 1,0

• Order of Elimination: T, R, B, C → (M,L)

• Order of Elimination: B, L, C, T → (M,R)
• Order of Elimination: T, C, R → (M,L) or (B,L)
• Does there exist a WDSE? No

15
Pure Strategy Nash Equilibrium

∗
A strategy profile s ∗ = (si∗ , s−i ) is called a pure strategy Nash equilibrium
(PSNE) if ∀i ∈ N, ∀si ∈ Si , the following is satisfied:

ui (si∗ , s−i
∗ ∗
) ≥ ui (si , s−i )

16
Pure Strategy Nash Equilibrium

∗
A strategy profile s ∗ = (si∗ , s−i ) is called a pure strategy Nash equilibrium
(PSNE) if ∀i ∈ N, ∀si ∈ Si , the following is satisfied:

ui (si∗ , s−i
∗ ∗
) ≥ ui (si , s−i )

No player has an incentive to unilaterally deviate

16
Pure Strategy Nash Equilibrium

∗
A strategy profile s ∗ = (si∗ , s−i ) is called a pure strategy Nash equilibrium
(PSNE) if ∀i ∈ N, ∀si ∈ Si , the following is satisfied:

ui (si∗ , s−i
∗ ∗
) ≥ ui (si , s−i )

No player has an incentive to unilaterally deviate

• If Iterated Elimination of Dominated Strategies leads to a single

strategy profile, it is a Nash Equilibrium

16
Pure Strategy Nash Equilibrium

∗
A strategy profile s ∗ = (si∗ , s−i ) is called a pure strategy Nash equilibrium
(PSNE) if ∀i ∈ N, ∀si ∈ Si , the following is satisfied:

ui (si∗ , s−i
∗ ∗
) ≥ ui (si , s−i )

No player has an incentive to unilaterally deviate

• If Iterated Elimination of Dominated Strategies leads to a single

strategy profile, it is a Nash Equilibrium (Converse not true)

16
Pure Strategy Nash Equilibrium

∗
A strategy profile s ∗ = (si∗ , s−i ) is called a pure strategy Nash equilibrium
(PSNE) if ∀i ∈ N, ∀si ∈ Si , the following is satisfied:

ui (si∗ , s−i
∗ ∗
) ≥ ui (si , s−i )

No player has an incentive to unilaterally deviate

• If Iterated Elimination of Dominated Strategies leads to a single

strategy profile, it is a Nash Equilibrium (Converse not true)
• Elimination of Strongly Dominated Strategies has no effect on a
game’s set of equilibria

16
Pure Strategy Nash Equilibrium

∗
A strategy profile s ∗ = (si∗ , s−i ) is called a pure strategy Nash equilibrium
(PSNE) if ∀i ∈ N, ∀si ∈ Si , the following is satisfied:

ui (si∗ , s−i
∗ ∗
) ≥ ui (si , s−i )

No player has an incentive to unilaterally deviate

• If Iterated Elimination of Dominated Strategies leads to a single

strategy profile, it is a Nash Equilibrium (Converse not true)
• Elimination of Strongly Dominated Strategies has no effect on a
game’s set of equilibria
• Elimination of Weakly Dominated Strategies may reduce the game’s
set of equilibria but will not create new set of equilibria

16
PSNE: Best Response Correspondence

Consider a function bi : S−i → 2Si defined as:

bi (s−i ) = {si ∈ Si : ui (si , s−i ) ≥ ui (si′ , s−i ) ∀si′ ∈ Si }

17
PSNE: Best Response Correspondence

Consider a function bi : S−i → 2Si defined as:

bi (s−i ) = {si ∈ Si : ui (si , s−i ) ≥ ui (si′ , s−i ) ∀si′ ∈ Si }

∗
(s1∗ , s2∗ , . . . , sn∗ ) is PSNE iff si∗ ∈ bi (s−i ) ∀i ∈ N

17
PSNE: Best Response Correspondence

Consider a function bi : S−i → 2Si defined as:

bi (s−i ) = {si ∈ Si : ui (si , s−i ) ≥ ui (si′ , s−i ) ∀si′ ∈ Si }

∗
(s1∗ , s2∗ , . . . , sn∗ ) is PSNE iff si∗ ∈ bi (s−i ) ∀i ∈ N

L C R
T 1,2 2,3 0,3
M 2,2 2,1 3,2
B 2,1 0,0 1,0

Determine all PSNE of this game

17
Example: Bach or Stravinsky

B S
B 2,1 0,0
S 0,0 1,2

18
Example: Co-ordination Game

Movie Study
Movie 10,10 0,0
Study 0,0 100,100

19
Example: Co-ordination Game

Movie Study
Movie 10,10 0,0
Study 0,0 100,100

• Two PSNE
• Which one to choose?

19
Example: Cournot Duopoly Competition

• Two Manufacturers 1 and 2 produce same product and compete for

the same market

• They select the quantities to produce: s1 , s2 ∈ [0, 1]

• When supply is s1 + s2 , the price is: 2 − s1 − s2

• Per unit manufacturing costs for the manufacturers are: c1 and c2

• What are the values of s1 and s2 in Nash equilibrium?

20
Example: Braess’ Paradox

• In a road network, 1000 vehicles want to go from node S to node T

• Edge weights denote the cost of traveling

• m = no. of vehicles passing through that edge

• Each player’s strategy set is {A, B} (routes S-A-T and S-B-T)

• Determine the PSNE and the traveling cost incurred by the players

21
Example: Braess’ Paradox

• Modify the road network by adding a superfast link from A to B

• Now, each player’s strategy set is {A, B, AB} (routes S-A-T, S-B-T,
and S-A-B-T)

• Determine the PSNE and the traveling cost incurred by the players

22
Example: Braess’ Paradox

• Modify the road network by adding a superfast link from A to B

• Now, each player’s strategy set is {A, B, AB} (routes S-A-T, S-B-T,
and S-A-B-T)

• Determine the PSNE and the traveling cost incurred by the players

• Did the players’ traveling cost reduce after adding the superfast link?

22
Classroom Game

23
Classroom Game

1
It’s Auction Time!

1 Image
24
sources: Pinterest, iStock, AnimationXpress
Classroom Game: Second Price Auction

• An indivisible object is offered for sale

25
Classroom Game: Second Price Auction

• An indivisible object is offered for sale

• The set of bidders is N with n bidders in the set

25
Classroom Game: Second Price Auction

• An indivisible object is offered for sale

• The set of bidders is N with n bidders in the set
• Let vi be the private value of bidder i for the object

25
Classroom Game: Second Price Auction

• An indivisible object is offered for sale

• The set of bidders is N with n bidders in the set
• Let vi be the private value of bidder i for the object
• Each bidder i submits a bid bi for the object in a sealed envelope

25
Classroom Game: Second Price Auction

• An indivisible object is offered for sale

• The set of bidders is N with n bidders in the set
• Let vi be the private value of bidder i for the object
• Each bidder i submits a bid bi for the object in a sealed envelope
• The winner of the object is the bidder with the highest bid

25
Classroom Game: Second Price Auction

• An indivisible object is offered for sale

• The set of bidders is N with n bidders in the set
• Let vi be the private value of bidder i for the object
• Each bidder i submits a bid bi for the object in a sealed envelope
• The winner of the object is the bidder with the highest bid
• If multiple bidders bid the highest, then a random bidder among
them is chosen as the winner (uniformly at random)

25
Classroom Game: Second Price Auction

• An indivisible object is offered for sale

• The set of bidders is N with n bidders in the set
• Let vi be the private value of bidder i for the object
• Each bidder i submits a bid bi for the object in a sealed envelope
• The winner of the object is the bidder with the highest bid
• If multiple bidders bid the highest, then a random bidder among
them is chosen as the winner (uniformly at random)
• The winner pays the second highest bid amount (or rather the
highest losing bid amount) to the auctioneer

25
Classroom Game: Second Price Auction

https://forms.gle/9vjxi3rjrMPwAsGh7

26
Second Price Auction: Game Formulation

• N: Set of n bidders
• Si : Any bid bi (say, in [0, 1])
• Utility function:

0, if bi < maxj∈N bj .
ui (bi , b−i ) = vi −maxj̸=i bj
|{k:bk =maxj∈N bj }| , otherwise.


27
Second Price Auction: Game Formulation

• N: Set of n bidders
• Si : Any bid bi (say, in [0, 1])
• Utility function:

0, if bi < maxj∈N bj .
ui (bi , b−i ) = vi −maxj̸=i bj
|{k:bk =maxj∈N bj }| , otherwise.


• What will rational and intelligent bidders do?

27
Second Price Auction: Result

The strategy bi = vi weakly dominates all other strategies, thus, every

bidder will bid their true valuations
This results in truthful elicitation

28