In the opinion of many serious players, the first great book on poker was ``Super/System: A Course in Power Poker'', produced in 1978 by Doyle Brunson, in conjunction with many of the top professional players of the day [27]. Each chapter addresses a different game, and is written by one of the foremost experts in the field. Despite the long history of poker literature, many of the most important concepts of poker strategy appeared in print for the first time in this book. While somewhat dated, it is still a valuable reference to players and researchers alike, because of the breadth of topics and poker formats discussed.
The opening chapter on general poker strategy was written by Brunson and discusses his philosophy and approach to the game. He introduces many of the topics which are to be discussed throughout the book, as well as some more nebulous views that would not fit elsewhere in the tome. While his opinions may or may not have any particular relevance to how a computer should play the game of poker, it is nevertheless important to determine, in theoretical terms, why a particular approach is highly successful against human opponents, while another is less suitable.
The key difference between Brunson's style and that of many theoreticians is his aggressiveness. While other authors suggest that aggressiveness is an important trait of top level players, the true value of this characteristic is belied by their conservative recommendations. Brunson makes it clear that his success is directly attributable to his willingness to take risks - betting or raising where other players would prefer a safer option.
It is now known that there are solid theoretical explanations for why a ``fast'' aggressive style is superior to a more conservative approach, but these observations have never been properly accounted for in previous studies.
Game theoretic studies generally have not considered situations where there is to be a draw of cards followed by another betting round. One of the consequences of this characteristic of real poker is that there is considerable risk involved in not betting when one has the best hand. Moreover, from a practical point of view, an opponent is far more likely to commit a misplay (such as folding a stronger hand) when confronted with a bet or raise, rather than a check or call. This suggests that an aggressive style is bound to be a more successful strategy than a more conservative approach, in practice.
Super/System provides a vast amount of poker knowledge, but understanding the value and necessity of aggressiveness is perhaps the book's most important contribution to poker praxis.
The first chapter dedicated to a specific game addresses Five-Card-Draw, and is written by Mike Caro, who is recognized as one of the foremost authors on the game of poker. In addition to sound basic strategy, Caro discusses other ways in which a strong player can gain an advantage over the opponents. One of these is through the use of ``tells'' - mannerisms exhibited by other players at the table which reveal information about their hands. Caro is considered to be one of the world's premier players at the art of poker psychology and behaviour, and later authored ``The Book of Tells'', in which he uses numerous photographs to illustrate the traits he considers to be valuable clues. Unfortunately, a computer algorithm is unlikely to be able to obtain or utilize such information, but neither is it vulnerable to such an attack.
The next game to be looked at is Seven-Card-Stud - one of the most enduring and challenging of all poker variations. The author is David ``Chip'' Reese, who is widely acknowledged to be among the best players in the world. Among the advanced concepts he addresses are the effect of the ante size, obtaining free cards, betting or check-raising a mediocre hand to narrow the field, the strength of concealed pairs, and adapting to unusual game conditions. This excellent treatment of Stud should be a definite asset to future researchers of this important poker variation.
The next chapter, authored by Joey Hawthorne, deals with three popular forms of Lowball poker, where the objective is to make the weakest possible hand. Some practical advice is given which may be useful to the casual player, but this section is not considered to be as strong as the other chapters of the book.
The fourth specialized chapter investigates High-Low Seven-Card-Stud, and is written by David Sklansky. Unfortunately, the particular game he addresses is no longer very popular, having given way to a variant where an Eight low or better is required to win the low half of the pot. Despite this fact, most of the information in this chapter remains indispensable, and was the first complete exposition of the special theoretical properties of high-low poker. Written in his usual clear and concise style, Sklansky demonstrates the unexpectedly high value of ``scooping'' (winning both halves of the pot), which thereby dictates all aspects of high-low strategy. Another general concept is the computation of the ``escape probability'', which determines the odds necessary to call with a weak hand in the hope of winning one half of the pot. Each element of the game (with no qualifier for low) is taken in turn, from playable starting hands to common playing errors. The chapter ends with a side discussion of declaration strategy, although this is not used in standard games, which are played ``cards speak'' (no declaration).
The first of two chapters on Texas Hold'em is written by Bobby Baldwin, who was one of the youngest players ever to win the world championship. This chapter investigates the limit poker form of Hold'em. While it has since been superseded by ``Texas Hold'em for the Advanced Player'', this chapter still provides a useful introduction to the game, without having to invest as much time and effort. It also complements the work of Sklansky and Malmuth, nicely illustrating some of the finer points of the game, including the varied play of drawing hands.
The final chapter is on No-Limit Hold'em, written by Doyle Brunson. No-Limit is a form of poker where a player may bet any amount at any time, up to and including the entire stack of chips he has on the table (called ``table stakes''). Pot-Limit is another form of ``big bet'' poker, which is quite similar in nature to no-limit, since it is usually appropriate to choose a bet of about the size pot.
No-Limit poker is strategically very different from the more common limit format. Since a player's whole stack is always in jeopardy, the risks and the rewards are much greater in no-limit poker. While limit poker is a patient exploitation of small advantages, no-limit emphasizes ``going for the kill'' (trapping an opponent for their whole stack). To be played at the level of the best humans, no-limit also requires a greater knowledge of the opponent, better judgement, and more courage. Because of the emphasis on these more ``human'' characteristics, limit poker may be more suitable to early studies in computer game playing. However, this does not suggest that computers will never be able to compete at the highest level in no-limit poker. On the contrary, a risk-neutral algorithm that is incapable of fear may eventually have a decided advantage over human opponents.
Because of the decreased importance of the technical aspects of the game, much of the chapter on no-limit Hold'em is spent discussing the appropriate mental attitude a player should adopt. While this may not be of much use to a mathematically oriented researcher, it does provide a valuable counter-point to be considered, and put into perspective.
After an extensive glossary, the book concludes with an appendix of poker probabilities and statistics. These tables, computed by Mike Caro, go far beyond the usual fare, exploring many interesting questions of practical value. Included are many game scenarios for Draw poker (with and without the Joker used as a limited wild card), Hold'em (starting hands, connecting with the board, long shots), Seven-Card-Stud (starting hands, chance of improving), and all forms of Lowball (with and without the Joker wild card). These tables form the basis of correct strategy, and are commendable both for their relevance and correctness.