Tartanian7: A Champion Two-Player No-Limit Texas Hold’em Poker-Playing Program

Authors: Noam Brown, Sam Ganzfried, Tuomas Sandholm

AAAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We experimented on the version of 2-player no-limit Texas Hold em (NLTH) used in the AAAI Annual Computer Poker Competition (ACPC). ... The results from the 2014 ACPC against all opponents are shown in Table 1.
Researcher Affiliation Academia Noam Brown, Sam Ganzfried, and Tuomas Sandholm Computer Science Department Carnegie Mellon University {nbrown, sganzfri, sandholm}@cs.cmu.edu
Pseudocode Yes Algorithm 1 Main abstraction algorithm Inputs: number of rounds R; round where public information abstraction is desired ˆr; number of public buckets C; number of desired private buckets per public bucket at round r, Br; abstraction algorithm used for round r, Ar for r = 1 to ˆr 1 do cluster information states at round r using Ar cluster public information states at round ˆr into C buckets (using our custom distance function and clustering algorithm) for r = ˆr to R do for c = 1 to C do cluster private information states at round r that have public information in public bucket c into Br buckets using abstraction algorithm Ar
Open Source Code No The paper does not provide any concrete access to source code for the methodology described.
Open Datasets No The paper mentions using data from the 'AAAI Annual Computer Poker Competition (ACPC)' but does not provide concrete access information (link, DOI, specific citation to a dataset paper) for a publicly available dataset.
Dataset Splits No The paper does not provide specific dataset split information (percentages, sample counts, or detailed methodology) for training, validation, or testing.
Hardware Specification Yes We ran our equilibrium-finding algorithm for 1,200 hours on a supercomputer (Blacklight) with a high inter-blade memory access latency using 961 cores (60 blades of 16 cores each, plus one core for the head blade), for a total of 1,153,200 core hours. Each blade had 128 GB RAM.
Software Dependencies No The paper does not provide specific software dependencies with version numbers.
Experiment Setup Yes We ran our equilibrium-finding algorithm for 1,200 hours on a supercomputer (Blacklight) with a high inter-blade memory access latency using 961 cores (60 blades of 16 cores each, plus one core for the head blade), for a total of 1,153,200 core hours. ... We compared using no threshold, purification, a threshold of 0.15, and using the new technique with a threshold of 0.2.