Advances in Computer Games: Many Games, Many Challenges

This book is the tenth in a well‐established series of conference proceedings, originally describing the progress of computer‐chess research. This particular conference was held in Graz, Austria, in late 2003, with participants from Europe, Japan, the US and Canada. However, after the ninth conference proceedings, an increasing number of paper submissions summarising research into non‐chess games led to the change in title and scope from chess to wider rule‐based computational games.

The 24 papers in these proceedings cover a wide number of games. However, Chess and Go are still predominant, with six papers each. Of the rest, Checkers is the focus of two papers, as is the game “Lines of Action”, with the remaining eight papers focusing on one game each: Hex, Othello, Amazons, Bao, Kriegspiel, Gaps, Oshi‐Zumo and New Wythoff.

The papers cover a very wide variety of computation techniques and analysis. The proceedings therefore are useful to people researching the creation of computer‐based solutions to games, and also to newer people to the field trying to grasp some concept of the scope of research development. For example, the six chess papers concern aspects of evaluation functions, pruning of the search, search and knowledge techniques, pattern recognition, modelling and strategies.

The two papers on Checkers are interesting as they demonstrate that, even with contemporary computer power, endgame databases can still be more effective than brute force “number crunching” when evaluating positions.

Some of the papers on the lesser‐known games are more fascinating to people less familiar with the field, due to the explanations of the rules and gameplay. One such paper, “Opponent‐model searching in Bao: conditions for a successful applicant”, concentrates on a game played predominantly in Tanzania and Zanzibar. In this game, the players “sow” stones in a matrix of holes on a board. By changes in sowing techniques and direction of sowing around the board, each player attempts to immobilise the other, thus winning the game. After explaining the rules of the game, the paper looks at opponent‐model searches in computerised simulations of the game, concluding that though adequate opponent prediction is useful in generating good results, extended search depth of possible outcomes is a more essential factor.

The papers and book assume a good knowledge of each game, and the standard game notations used. Set theory is used in several papers, as are calculus and other advanced mathematical tools. I would recommend these proceedings to anyone with an interest in the field who possess knowledge of these games, as well as the requisite computing and mathematical knowledges.

Editor's note. For further information on the connection between computer games and LIS research see Kirriemuir (2000, 2001).