
Economics Economics Forum  Financial Mathematics, Econometrics, Operations Research, Mathematical Finance, Computational Finance 
 LinkBack  Thread Tools  Display Modes 
April 22nd, 2009, 07:10 AM  #1 
Newbie Joined: Apr 2009 Posts: 1 Thanks: 0  [game theory] Election game The game is: N players ("candidates") make their moves sequentially. The move is a decision to play or not to play, and if a player chooses to play, then he must take a real number from [0;1] (that would be candidate's "political views", 0 might mean "democrat" and 1  "republican"). Every player knows all moves of previous players, but does not know moves of next players (e.g. the first player knows no moves, the last one  knows all moves). After all moves are made, votes are counted: number of votes for a candidate X equals the length of a segment, where all points are closer to the number proposed by X on his move, than to numbers, proposed by any other candidate. If k players have chosen the same number, than the length of correspondent segment should be divided by k. The candidate, who has the greatest number of "votes" is the winner and gets a prize = $1. All other players, who played the game, but lost, bear costs of $1 (i.e.their prize=1). The players, who chose not to play, get $0 (no prize, no costs). Again, in case k players have got the same number of votes, the prize will be divided between these players equally: each player"semifinalist" will get $(1/k). The goal is of course to win maximum amount of money. The hypothesis:subgameperfect Nash equilibrium for this game is: the first player choses the point 0.5 , all other players except the last chose not to play, and the last player also choses the point 0.5. The problem: This statement is true for 3 players. I see no reason it will not be true for any finite number of players, but I have no idea how to prove that, so I will be grateful for any help, interesting suggestions/ideas or similar problems and literature relevant to the question. Thanks in advance 
April 22nd, 2009, 11:35 AM  #2 
Global Moderator Joined: Nov 2006 From: UTC 5 Posts: 16,046 Thanks: 938 Math Focus: Number theory, computational mathematics, combinatorics, FOM, symbolic logic, TCS, algorithms  Re: [game theory] Election game
You may consider Black's theorem to be related: Duncan Black, "On the Rationale of Group Decisionmaking", Journal of Political Economy 56 (194, pp. 2334. 

Tags 
election, game, theory 
Thread Tools  
Display Modes  

Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
game theory  blabla  Economics  1  January 19th, 2014 03:53 AM 
Game theory  gaussrelatz  New Users  2  January 18th, 2014 06:43 AM 
Analyzing a game for fairness? (game theory question)  musicgold  Economics  0  March 14th, 2012 08:04 PM 
[game theory] Election game  lkzavr  Applied Math  1  December 31st, 1969 04:00 PM 
game theory  blabla  Applied Math  0  December 31st, 1969 04:00 PM 