Several resources other than the written word are available to researchers interested in computer poker. Many of these are accessible through the Internet, using the World Wide Web (WWW), anonymous ftp, or various other protocols.
The Usenet newsgroup ``rec.gambling.poker'' is dedicated to the analysis and discussion of a wide variety of poker topics. Many authorities on the game are among the regular contributors to this newsgroup, including well-known authors, theoreticians, casino directors, and numerous professional and expert players. The majority of articles posted to this forum are interesting and well thought out, resulting in a much higher 'signal-to-noise' ratio than is typical for most newsgroups.
Archive sites for this newsgroup are available for the perusal of past articles [126, 127]. A ``Frequently Asked Questions'' (FAQ) file covers a wide variety of common topics, including the rules of Hold'em and other variations, good books and magazines, poker probability questions, and an extensive glossary of poker terminology [128]. A WWW home page for rec.gambling.poker also serves as a starting point for locating many related Internet sites [129].
Of particular interest is a computer server which enables anyone to play poker against other people from around the world. This program exists on a isolated branch of the Internet Relay Chat (IRC) system, and features a variety of poker games and players at all skill levels [124].
Residing on this server are a number of ``bots'' - automated programs which facilitate the play of games, analyze the probabilistic outcomes of each situation, and record a summary of each hand played [124, 121, 122] . The information acquired from this observation is compiled into databases, which provide a valuable and unique source of raw data on poker playing [116]. Michael Maurer, the author and maintainer of the observer programs and IRC poker databases, has processed some of this information in a number of different ways [118, 119, 120]. For example, this knowledge was used to conduct a study of mathematical variance in poker games, varying in skill level, number of players, and player styles [123]. Such a complete study has never before been possible, and many other inquiries could conceivably be addressed with this body of information.
The IRC server is also amenable to hosting poker playing programs - allowing them to compete against human opponents or other programs with a minimal amount of supervision. This could serve as a useful testing ground during the development of poker algorithms, and as a battleground for competition between rival programs.
In the spirit of ``Percentage Hold'em'' [28], Michael Maurer has also conducted simulation experiments for the evaluation of starting hands for several other games, including Hi-Low Hold'em and Hi-Low Seven-Card-Stud [111]. His WWW poker page contains many other interesting and useful features [117].
Some high quality software support is also freely available via the Internet. An extremely fast winning hand evaluator written in C is available for Unix systems, and could serve at the core of higher level algorithms [112, 113]. Superseding some of the ``Percentage Hold'em'' information, programs are now available which can compute exact values for head-to-head match-ups of Hold'em starting hands. These programs use highly optimized hand comparators, and determine all possible outcomes by exhaustive enumeration [114]. Simulator programs for comparing hands in Hold'em and Seven-Card Stud are also available [125].
There are several commercial poker programs available on the market [115]. The level of play demonstrated by these programs is generally quite poor, although some modest improvements have been attained over the past few years. Unfortunately, the proprietary nature of these products makes them of limited use to academic researchers. Moreover, it appears that the business oriented developers of these programs devote much more effort to enhancing certain selling features, such as attractive graphics and user interfaces, rather than the greater challenge of improving the playing strength.
Some free demonstration versions of commercial programs can be downloaded from Internet sites, including a fully functional and relatively strong program for Seven-Card-Stud [109].
Finally, one of the most invaluable resources to the poker researcher is human expert consultation. Authoring a program that exhibits superior skill will likely involve solving a myriad of practical problems. The most effective means of attaining the the insight necessary to overcome these obstacles is through personal mastery of the game. For this reason, a complete understanding of the theoretical principles and strategic dynamics of the game is an essential quality for pioneers in poker research. Fortunately, much of this ability can be achieved fairly quickly, even by a newcomer to poker, through the study of the classic books reviewed in this chapter, and with regular practice.