Skip to content

Latest commit

 

History

History
21 lines (17 loc) · 1.16 KB

README.md

File metadata and controls

21 lines (17 loc) · 1.16 KB

Optimal Poker Strategies Under Collusion

This is the data collection code I wrote for my thesis on Simulating Fictitious Play in Simplified Poker Variants.

This code was a means to an end -- noteably to compute a Nash Equilibrium in poker-like games for my thesis. The code is not well documented and although it provides a good enough abstraction for specifying a wide variety of poker-like games, it is not particularly performant.

Usage

You can see a few of the computations I performed in run.sh, which runs ComputeNash.java.
The main() is currently hard-coded to find a Nash Equilibrium strategy in 3-player Mercer Holdem -- a simple poker variant. Running

java ComputeNash [pos1] [pos2] [tolerance] [output prefix]

computes both a non-collusion and collusion Nash Equilibrium (where pos1 and pos2 are the zero-indexed positions of the colluding players) of the game via ficticious play.

Read the paper for more information on the structure of the program and analysis of the results.