Download the full-sized PDF of A Tool for the Direct Assessment of Poker DecisionsDownload the full-sized PDF



Permanent link (DOI):


Export to: EndNote  |  Zotero  |  Mendeley


This file is in the following communities:

Computing Science, Department of


This file is in the following collections:

Technical Reports (Computing Science)

A Tool for the Direct Assessment of Poker Decisions Open Access


Author or creator
Billings, Darse
Kan, Morgan
Additional contributors
poker DIVAT analysis tool
Computer Games
Type of item
Computing Science Technical Report
Computing science technical report ID
Technical report TR06-07. The development of the Ignorant Value Assessment Tool (DIVAT) for two-player Limit Texas Hold'em is discussed in detail. The tool is then applied to several poker matches to obtain a statistically unbiased reduced-variance analysis of skill differences. Please note: this technical report is currently the final version submitted for the ICGA journal article. The authors plan to update this technical report in the coming months, showing other features of the tool, and providing detailed analyses of real matches played in the Vegas'05 and AAAI'06 competitions. Abstract to the ICGA journal submission: The element of luck permeates every aspect of the game of poker. Random stochastic outcomes introduce a large amount of noise that can make it very difficult to distinguish a good player from a bad one, much less trying to quantify small differences in skill. Good methods for directly assessing the quality of decisions exist for other stochastic games, including backgammon and blackjack. However, those are perfect information games, where the notion of an objectively best move is well-defined, which unfortunately is not the case for imperfect information games, in general. The Ignorant Value Assessment Tool, DIVAT, uses perfect knowledge (hindsight) analysis to quantify the value of each player decision made in a game of two-player Limit Texas Hold'em. Comparisons are made against a realistic baseline betting sequence, which is based on quasi-equilibrium policies and game-theoretic invariant frequencies for raising and folding. Much of the relevant context involved in each decision is simply ignored, out of necessity; but enough is retained to provide a reasonably accurate yardstick, which is then applied equally to all players. The frequency and magnitude of expected value differences between player actions, relative to the baseline, provides a statistically unbiased low-variance estimate of the long-term expected outcome between the players.
Date created
License information
Creative Commons Attribution 3.0 Unported

Citation for previous publication

Link to related item

File Details

Date Uploaded
Date Modified
Audit Status
Audits have not yet been run on this file.
File format: pdf (Portable Document Format)
Mime type: application/pdf
File size: 3778162
Last modified: 2015:10:12 17:33:13-06:00
Filename: TR06-07.pdf
Original checksum: 7f0bc6ecde05e546036333615364049b
Well formed: true
Valid: true
Page count: 25
Activity of users you follow
User Activity Date