Chess Skill in Man and Machine (Texts and Monographs in Computer Science)

Rating:
Summary: One of the rare books on chess programming
Review: Computer chess programming has always interested me, so I decided to document me on the subject and with a little luck, to develop a simple chess playing program.

"Chess skill in man and machine" was unfortunately not the right answer. What this book really is? It's a description of the advancement in computer chess. It contains the move list and comments of over 40 computer chess games (that which I find completely unnecessary), plus descriptions of how famous chess programs were created (again, some of them unnecessary and certainly repetive). In overall, only chapter 2 and 3 deserved my attention. They describe the basics (underline "basics") of chess programming such as board representation, position evaluation and searching.

Did I mention that the book uses the matrix chess notation (ex: P-QB4), something that even as an active chess player, I was unable to understand without further research?

Rating:
Summary: One of the best books on the foundations of computer chess
Review: This book is one of the pioneer works on the subject, as its publication date shows clearly. Despite its age, it remains as one of the most fascinating introductions to computer chess, and most of the ideas it presents are still valid. Its multiple authors cover all aspects of chess playing, from technical expositions of some of the best programs of that time, to physiological and psychological considerations.

In "A brief history of computer chess tournaments: 1970-1975", we are introduced to the atmosphere of the early tournaments, the diverse friendly matches between US and USSR chess computers, and several US and international championships, with many of the most interesting games fully commented and analyzed.

The next chapter, "Human chess skill" focuses in how does a human player select a move in the game of chess, the role of perception, the search mechanism, visualization, as well as other tipically human aspects such as motivation. Several tests applied to human players ranging from novices to grandmasters are presented and discussed.

After that introspective look at we humans, and our not-so-well understood thought processes, "An introduction to computer chess" begin to shift the focus to the computer, including such basic topics as how to represent the chess board, the moves, the status, how to generate the legal moves, search strategies, position evaluation, so that by the end of the chapter, all necessary foundations are well stablished for the rest of the book.

With Chapter 4, "Chess 4.5 - The Northwestern University chess progam" we begin the most technical part of the book. Here, authors David J. Slate and Lawrence R. Atkin show us with great style the internal workings of their famous chess program, many times world champion, and the one mostly used against IM David Levy for the famous Levy's bet. The details are sufficient to help a lot anyone contemplating the possibility of writing his/her own chess program. Modestly, the authors assume the limitations of their creature, and offer good advice on how it can be incrementally improved.

Chapter 5, "PEASANT: An endgame program for kings and pawns" provides yet another close scrutiny of a chess program, though this time with the important novelty that it is an specialized chess program, one specifically designed for a certain class of very frequent endgames. Monroe Newborn, its author, fully describes the inner workings, and most importantly, produces a set of tests for his program, with commented results.

The next chapter, "Plans, goals, and search strategies for the selection of a move in chess" tries to center on how do human players select good chess moves when having just a few seconds to consider the position (i.e: blitz chess), and then introduces a chess program specifically designed to play speed chess, without recourse to tree searching. This quite intriguing approach more closely mimics the human behaviour, to the point of even producing the same kind of erroneous moves a human player would play at blitz speeds.

As an alternative to the standard alpha-beta search techniques, Larry R. Harris introduces us to "The heuristic search: An alternative to the alpha-beta minimax procedure", where it presents what it considers important pitfalls of that search strategy, fully commented with specific examples, and proposes a new paradigm that addresses each and everyone of them from the start, thus truly directing the search in an intelligent way, as opposed to brute force, so that each aspect of the position can be ascertained as soon as possible, before going to other places in the search tree.

After these mostly technical chapters, in "Man and machine: Chess achievements and chess thinking", professor Eliot Hearst, a member of the Psychology Department at Indiana University, evaluates the present status of computer chess from the perspective of someone very knowledgeable with the game, as he is a rather skilled chess player and columnist. He includes many good practical examples, to make his points even clearer.

The book closes with a number of games played by Chess 4.5 and 4.6 in competitions during 1976, 1977, and 1978, that show a remarkable improvement on the rather pessimistic forecastings most experts agreed upon at that time.

Rating:
Summary: An excellent historical reference.
Review: This book shows the state of the art at the end of the 1970's. Though there have been huge changes since then, and tremendous gaps in our understanding have been filled, I still can't recommend this book highly enough. This book was published at the time when the best programs changed over from selective search to brute force. Nowadays we know that brute force is the way to go, but at that time even programmers who were winning tournaments using brute force techniques had little faith in their ultimate viability. The authors's speculation about the roles of search and evaluation is very interesting from the historical perspective.

One chapter of this book is worth the entire price. Slate and Atkins describe Chess 4.5 in one chapter. That chapter remains to this day the best description of an "attack-table" chess engine ever written, though you will need some additional reading to create a modern program on that basis.