Upload
others
View
7
Download
0
Embed Size (px)
Citation preview
Review
We considered the game:
I There is a spectrum of 10 points on a certain political issue
I There are two candidates
I 10% of the voters hold each position
I Voters will vote for the candidate who holds the closest views
I Candidates will split the vote of views that are the samedistance to both candidates
I Each candidate wants to maximize their share of the vote
Review
We considered the game:
I There is a spectrum of 10 points on a certain political issue
I There are two candidates
I 10% of the voters hold each position
I Voters will vote for the candidate who holds the closest views
I Candidates will split the vote of views that are the samedistance to both candidates
I Each candidate wants to maximize their share of the vote
Review
We considered the game:
I There is a spectrum of 10 points on a certain political issue
I There are two candidates
I 10% of the voters hold each position
I Voters will vote for the candidate who holds the closest views
I Candidates will split the vote of views that are the samedistance to both candidates
I Each candidate wants to maximize their share of the vote
Review
We considered the game:
I There is a spectrum of 10 points on a certain political issue
I There are two candidates
I 10% of the voters hold each position
I Voters will vote for the candidate who holds the closest views
I Candidates will split the vote of views that are the samedistance to both candidates
I Each candidate wants to maximize their share of the vote
Review
We considered the game:
I There is a spectrum of 10 points on a certain political issue
I There are two candidates
I 10% of the voters hold each position
I Voters will vote for the candidate who holds the closest views
I Candidates will split the vote of views that are the samedistance to both candidates
I Each candidate wants to maximize their share of the vote
Review
We considered the game:
I There is a spectrum of 10 points on a certain political issue
I There are two candidates
I 10% of the voters hold each position
I Voters will vote for the candidate who holds the closest views
I Candidates will split the vote of views that are the samedistance to both candidates
I Each candidate wants to maximize their share of the vote
Political Spectrum
I Are there any dominated strategies?
I 1 is weakly dominated by 2I 10 is weakly dominated by 9I 3 does not dominate 2
but after we remove 1 it does
I If we iterate this, the candidates end up in the centralpositions
I This is The Median Voter Theorem“Majority rule voting will select the median preference”
Political Spectrum
I Are there any dominated strategies?I 1 is weakly dominated by 2
I 10 is weakly dominated by 9I 3 does not dominate 2
but after we remove 1 it does
I If we iterate this, the candidates end up in the centralpositions
I This is The Median Voter Theorem“Majority rule voting will select the median preference”
Political Spectrum
I Are there any dominated strategies?I 1 is weakly dominated by 2I 10 is weakly dominated by 9
I 3 does not dominate 2
but after we remove 1 it does
I If we iterate this, the candidates end up in the centralpositions
I This is The Median Voter Theorem“Majority rule voting will select the median preference”
Political Spectrum
I Are there any dominated strategies?I 1 is weakly dominated by 2I 10 is weakly dominated by 9I 3 does not dominate 2
but after we remove 1 it does
I If we iterate this, the candidates end up in the centralpositions
I This is The Median Voter Theorem“Majority rule voting will select the median preference”
Political Spectrum
I Are there any dominated strategies?I 1 is weakly dominated by 2I 10 is weakly dominated by 9I 3 does not dominate 2
but after we remove 1 it does
I If we iterate this, the candidates end up in the centralpositions
I This is The Median Voter Theorem“Majority rule voting will select the median preference”
Political Spectrum
I Are there any dominated strategies?I 1 is weakly dominated by 2I 10 is weakly dominated by 9I 3 does not dominate 2
but after we remove 1 it does
I If we iterate this, the candidates end up in the centralpositions
I This is The Median Voter Theorem“Majority rule voting will select the median preference”
Political Spectrum
I Are there any dominated strategies?I 1 is weakly dominated by 2I 10 is weakly dominated by 9I 3 does not dominate 2
but after we remove 1 it does
I If we iterate this, the candidates end up in the centralpositions
I This is The Median Voter Theorem“Majority rule voting will select the median preference”
Median Voter Theorem
Problems?
I Assumed distribution was constant
I Assuming full voter turnout
I Assuming that there are only two candidates
I Assuming voters are rational
I Assuming that candidates are rational, and that they assumethat there opponent is rational
Examples:
I Kennedy (‘60)
I Nixon (‘68)
I Clinton (‘92)
I Affordable Care Act
I Gas station distribution
Median Voter Theorem
Problems?
I Assumed distribution was constant
I Assuming full voter turnout
I Assuming that there are only two candidates
I Assuming voters are rational
I Assuming that candidates are rational, and that they assumethat there opponent is rational
Examples:
I Kennedy (‘60)
I Nixon (‘68)
I Clinton (‘92)
I Affordable Care Act
I Gas station distribution
Median Voter Theorem
Problems?
I Assumed distribution was constant
I Assuming full voter turnout
I Assuming that there are only two candidates
I Assuming voters are rational
I Assuming that candidates are rational, and that they assumethat there opponent is rational
Examples:
I Kennedy (‘60)
I Nixon (‘68)
I Clinton (‘92)
I Affordable Care Act
I Gas station distribution
Median Voter Theorem
Problems?
I Assumed distribution was constant
I Assuming full voter turnout
I Assuming that there are only two candidates
I Assuming voters are rational
I Assuming that candidates are rational, and that they assumethat there opponent is rational
Examples:
I Kennedy (‘60)
I Nixon (‘68)
I Clinton (‘92)
I Affordable Care Act
I Gas station distribution
Median Voter Theorem
Problems?
I Assumed distribution was constant
I Assuming full voter turnout
I Assuming that there are only two candidates
I Assuming voters are rational
I Assuming that candidates are rational, and that they assumethat there opponent is rational
Examples:
I Kennedy (‘60)
I Nixon (‘68)
I Clinton (‘92)
I Affordable Care Act
I Gas station distribution
Median Voter Theorem
Problems?
I Assumed distribution was constant
I Assuming full voter turnout
I Assuming that there are only two candidates
I Assuming voters are rational
I Assuming that candidates are rational, and that they assumethat there opponent is rational
Examples:
I Kennedy (‘60)
I Nixon (‘68)
I Clinton (‘92)
I Affordable Care Act
I Gas station distribution
Median Voter Theorem
Problems?
I Assumed distribution was constant
I Assuming full voter turnout
I Assuming that there are only two candidates
I Assuming voters are rational
I Assuming that candidates are rational, and that they assumethat there opponent is rational
Examples:
I Kennedy (‘60)
I Nixon (‘68)
I Clinton (‘92)
I Affordable Care Act
I Gas station distribution
Median Voter Theorem
Problems?
I Assumed distribution was constant
I Assuming full voter turnout
I Assuming that there are only two candidates
I Assuming voters are rational
I Assuming that candidates are rational, and that they assumethat there opponent is rational
Examples:
I Kennedy (‘60)
I Nixon (‘68)
I Clinton (‘92)
I Affordable Care Act
I Gas station distribution
Median Voter Theorem
Problems?
I Assumed distribution was constant
I Assuming full voter turnout
I Assuming that there are only two candidates
I Assuming voters are rational
I Assuming that candidates are rational, and that they assumethat there opponent is rational
Examples:
I Kennedy (‘60)
I Nixon (‘68)
I Clinton (‘92)
I Affordable Care Act
I Gas station distribution
CampingI Alex and Bob are going camping
I Alex wants to camp at a high altitudeI Bob wants to camp at a low altitude
Camping spots (with elevation in 1000s of feet):
4
6 1
4
5
4 3
1
5
2 3
3
7
2 2
2
I Alex chooses east-west stripI Bob chooses north-south stripI Says Alex’s payoff is the elevation, and Bob’s payoff is the
opposite
CampingI Alex and Bob are going camping
I Alex wants to camp at a high altitude
I Bob wants to camp at a low altitude
Camping spots (with elevation in 1000s of feet):
4
6 1
4
5
4 3
1
5
2 3
3
7
2 2
2
I Alex chooses east-west stripI Bob chooses north-south stripI Says Alex’s payoff is the elevation, and Bob’s payoff is the
opposite
CampingI Alex and Bob are going camping
I Alex wants to camp at a high altitudeI Bob wants to camp at a low altitude
Camping spots (with elevation in 1000s of feet):
4
6 1
4
5
4 3
1
5
2 3
3
7
2 2
2
I Alex chooses east-west stripI Bob chooses north-south stripI Says Alex’s payoff is the elevation, and Bob’s payoff is the
opposite
CampingI Alex and Bob are going camping
I Alex wants to camp at a high altitudeI Bob wants to camp at a low altitude
Camping spots (with elevation in 1000s of feet):
4
6 1
4
5
4 3
1
5
2 3
3
7
2 2
2
I Alex chooses east-west stripI Bob chooses north-south stripI Says Alex’s payoff is the elevation, and Bob’s payoff is the
opposite
CampingI Alex and Bob are going camping
I Alex wants to camp at a high altitudeI Bob wants to camp at a low altitude
Camping spots (with elevation in 1000s of feet):
4
6 1
4
5
4 3
1
5
2 3
3
7
2 2
2
I Alex chooses east-west stripI Bob chooses north-south stripI Says Alex’s payoff is the elevation, and Bob’s payoff is the
opposite
CampingI Alex and Bob are going camping
I Alex wants to camp at a high altitudeI Bob wants to camp at a low altitude
Camping spots (with elevation in 1000s of feet):
4
6 1
4
5
4 3
1
5
2 3
3
7
2 2
2
I Alex chooses east-west stripI Bob chooses north-south strip
I Says Alex’s payoff is the elevation, and Bob’s payoff is theopposite
CampingI Alex and Bob are going camping
I Alex wants to camp at a high altitudeI Bob wants to camp at a low altitude
Camping spots (with elevation in 1000s of feet):
4
6 1
4
5
4 3
1
5
2 3
3
7
2 2
2
I Alex chooses east-west stripI Bob chooses north-south stripI Says Alex’s payoff is the elevation, and Bob’s payoff is the
opposite
Camping
I Can rule out dominated strategies:
4
6 1
4
5
4 3
1
5
2 3
3
7
2 2
2
I Problem: we’re now stuck
I New idea: find points where no player regrets their choice
Camping
I Can rule out dominated strategies:
4
6 1
4
5
4 3
1
5
2 3
3
7
2 2
2
I Problem: we’re now stuck
I New idea: find points where no player regrets their choice
Camping
I Can rule out dominated strategies:
4
6 1
4
5
4 3
1
5
2 3
3
7
2 2
2
I Problem: we’re now stuck
I New idea: find points where no player regrets their choice
Camping
I Consider the following campsite:
4
6 1
4
5
4 3
1
5
2 3
3
7
2 2
2
I If they chose this spot, would either Alex or Bob have regrets?
I No
I Such an outcome is called a Nash equilibrium
I More formally, a strategy profile s1, . . . , sn is a Nashequilibrium if u(si , s−i ) ≥ u(s∗i , s−i ) for each i
I So, if all other players’ fix their strategy, you can’t do better
Camping
I Consider the following campsite:
4
6 1
4
5
4 3
1
5
2 3
3
7
2 2
2
I If they chose this spot, would either Alex or Bob have regrets?
I No
I Such an outcome is called a Nash equilibrium
I More formally, a strategy profile s1, . . . , sn is a Nashequilibrium if u(si , s−i ) ≥ u(s∗i , s−i ) for each i
I So, if all other players’ fix their strategy, you can’t do better
Camping
I Consider the following campsite:
4
6 1
4
5
4 3
1
5
2 3
3
7
2 2
2
I If they chose this spot, would either Alex or Bob have regrets?
I No
I Such an outcome is called a Nash equilibrium
I More formally, a strategy profile s1, . . . , sn is a Nashequilibrium if u(si , s−i ) ≥ u(s∗i , s−i ) for each i
I So, if all other players’ fix their strategy, you can’t do better
Camping
I Consider the following campsite:
4
6 1
4
5
4 3
1
5
2 3
3
7
2 2
2
I If they chose this spot, would either Alex or Bob have regrets?
I No
I Such an outcome is called a Nash equilibrium
I More formally, a strategy profile s1, . . . , sn is a Nashequilibrium if u(si , s−i ) ≥ u(s∗i , s−i ) for each i
I So, if all other players’ fix their strategy, you can’t do better
Camping
I Consider the following campsite:
4
6 1
4
5
4 3
1
5
2 3
3
7
2 2
2
I If they chose this spot, would either Alex or Bob have regrets?
I No
I Such an outcome is called a Nash equilibrium
I More formally, a strategy profile s1, . . . , sn is a Nashequilibrium if u(si , s−i ) ≥ u(s∗i , s−i ) for each i
I So, if all other players’ fix their strategy, you can’t do better
Camping
I Consider the following campsite:
4
6 1
4
5
4 3
1
5
2 3
3
7
2 2
2
I If they chose this spot, would either Alex or Bob have regrets?
I No
I Such an outcome is called a Nash equilibrium
I More formally, a strategy profile s1, . . . , sn is a Nashequilibrium if u(si , s−i ) ≥ u(s∗i , s−i ) for each i
I So, if all other players’ fix their strategy, you can’t do better
Camping
I How do we find Nash equilibria?
I For others’ strategies, determine your best strategyI See where these coincide for the players
4
6 1
4
5
4 3
1
5
2 3
3
7
2 2
2
3
Camping
I How do we find Nash equilibria?I For others’ strategies, determine your best strategyI See where these coincide for the players
4
6 1
4
5
4 3
1
5
2 3
3
7
2 2
2
3
Camping
I How do we find Nash equilibria?I For others’ strategies, determine your best strategyI See where these coincide for the players
4
6 1
4
5
4 3
1
5
2 3
3
7
2 2
2
3
Nash Equilibria
I See handout #5
I Note:
I There can be more than one Nash equilibriumI Nash equilibria are not always the best solutionsI Nash equilibria never lie on strictly dominated strategiesI They can lie on weakly dominated strategies
Nash Equilibria
I See handout #5I Note:
I There can be more than one Nash equilibriumI Nash equilibria are not always the best solutionsI Nash equilibria never lie on strictly dominated strategiesI They can lie on weakly dominated strategies
Nash Equilibria
I See handout #5I Note:
I There can be more than one Nash equilibrium
I Nash equilibria are not always the best solutionsI Nash equilibria never lie on strictly dominated strategiesI They can lie on weakly dominated strategies
Nash Equilibria
I See handout #5I Note:
I There can be more than one Nash equilibriumI Nash equilibria are not always the best solutions
I Nash equilibria never lie on strictly dominated strategiesI They can lie on weakly dominated strategies
Nash Equilibria
I See handout #5I Note:
I There can be more than one Nash equilibriumI Nash equilibria are not always the best solutionsI Nash equilibria never lie on strictly dominated strategies
I They can lie on weakly dominated strategies
Nash Equilibria
I See handout #5I Note:
I There can be more than one Nash equilibriumI Nash equilibria are not always the best solutionsI Nash equilibria never lie on strictly dominated strategiesI They can lie on weakly dominated strategies
The Investment Game
I You have a choice:
I You can invest $20I You can choose to not invest
I If more than 90% of the class chooses to invest, you earn $10on top of your original investment
I Otherwise, you lose your $20
I Choose whether or not you want to investI What are the Nash equilibria?
I We can find them by guessing and testingI All invest, or none invest
The Investment Game
I You have a choice:I You can invest $20I You can choose to not invest
I If more than 90% of the class chooses to invest, you earn $10on top of your original investment
I Otherwise, you lose your $20
I Choose whether or not you want to investI What are the Nash equilibria?
I We can find them by guessing and testingI All invest, or none invest
The Investment Game
I You have a choice:I You can invest $20I You can choose to not invest
I If more than 90% of the class chooses to invest, you earn $10on top of your original investment
I Otherwise, you lose your $20
I Choose whether or not you want to investI What are the Nash equilibria?
I We can find them by guessing and testingI All invest, or none invest
The Investment Game
I You have a choice:I You can invest $20I You can choose to not invest
I If more than 90% of the class chooses to invest, you earn $10on top of your original investment
I Otherwise, you lose your $20
I Choose whether or not you want to investI What are the Nash equilibria?
I We can find them by guessing and testingI All invest, or none invest
The Investment Game
I You have a choice:I You can invest $20I You can choose to not invest
I If more than 90% of the class chooses to invest, you earn $10on top of your original investment
I Otherwise, you lose your $20
I Choose whether or not you want to invest
I What are the Nash equilibria?
I We can find them by guessing and testingI All invest, or none invest
The Investment Game
I You have a choice:I You can invest $20I You can choose to not invest
I If more than 90% of the class chooses to invest, you earn $10on top of your original investment
I Otherwise, you lose your $20
I Choose whether or not you want to investI What are the Nash equilibria?
I We can find them by guessing and testingI All invest, or none invest
The Investment Game
I You have a choice:I You can invest $20I You can choose to not invest
I If more than 90% of the class chooses to invest, you earn $10on top of your original investment
I Otherwise, you lose your $20
I Choose whether or not you want to investI What are the Nash equilibria?
I We can find them by guessing and testing
I All invest, or none invest
The Investment Game
I You have a choice:I You can invest $20I You can choose to not invest
I If more than 90% of the class chooses to invest, you earn $10on top of your original investment
I Otherwise, you lose your $20
I Choose whether or not you want to investI What are the Nash equilibria?
I We can find them by guessing and testingI All invest, or none invest
The Investment Game
I Let’s play the game again.
I What happened to peoples’ strategies?I This is an example of a coordination game:
I There are multiple Nash equilibriaI Saying your strategy out loud is beneficial
I Other players will have no reason to think that you’re lyingI Other players will choose the corresponding equilibrium point
The Investment Game
I Let’s play the game again.
I What happened to peoples’ strategies?
I This is an example of a coordination game:
I There are multiple Nash equilibriaI Saying your strategy out loud is beneficial
I Other players will have no reason to think that you’re lyingI Other players will choose the corresponding equilibrium point
The Investment Game
I Let’s play the game again.
I What happened to peoples’ strategies?I This is an example of a coordination game:
I There are multiple Nash equilibriaI Saying your strategy out loud is beneficial
I Other players will have no reason to think that you’re lyingI Other players will choose the corresponding equilibrium point
The Investment Game
I Let’s play the game again.
I What happened to peoples’ strategies?I This is an example of a coordination game:
I There are multiple Nash equilibria
I Saying your strategy out loud is beneficial
I Other players will have no reason to think that you’re lyingI Other players will choose the corresponding equilibrium point
The Investment Game
I Let’s play the game again.
I What happened to peoples’ strategies?I This is an example of a coordination game:
I There are multiple Nash equilibriaI Saying your strategy out loud is beneficial
I Other players will have no reason to think that you’re lyingI Other players will choose the corresponding equilibrium point
The Investment Game
I Let’s play the game again.
I What happened to peoples’ strategies?I This is an example of a coordination game:
I There are multiple Nash equilibriaI Saying your strategy out loud is beneficial
I Other players will have no reason to think that you’re lying
I Other players will choose the corresponding equilibrium point
The Investment Game
I Let’s play the game again.
I What happened to peoples’ strategies?I This is an example of a coordination game:
I There are multiple Nash equilibriaI Saying your strategy out loud is beneficial
I Other players will have no reason to think that you’re lyingI Other players will choose the corresponding equilibrium point