ICCA Journal, Vol. 17, No. 1
Jos Uiterwijk
The ICCA Journal, Vol. 17, No. 1 (March 1994) has just been
released from the printer.
For those not knowing this Journal:
======================================================================
The ICCA Journal is the Journal of the ICCA, the International
Computer Chess Association. The ICCA Journal is devoted to
scientific publications (articles, research notes, book reviews),
reports of tournaments, announcements and other information on any
aspect of computer chess.
The ICCA Journal is published 4 times a year and a yearly subscription
costs fl 50.- (Dutch guilders) or $ 30.- (US dollars). More information
can be obtained by sending an Email message to
======================================================================
The present issue (Vol. 17, No. 1) contains among others:
Articles:
--------
Don Beal and Martin C. Smith: Random Evaluations in Chess.
The paper reports experiments using random numbers as "evaluations"
in chess. Although this results in random play with a depth-1
search, it is found that strength of play rises rapidly when the
depth of lookahead is increased. This counter-intuitive result
is discussed and its implications for game-playing are given.
Hiroyuki Iida, Jos Uiterwijk, Jaap van den Herik and Bob Herschberg:
Potential Applications of Opponent-Model Search. Part 2: risks and strategies.
This is the second part of some articles devoted to modelling an
opponent by assumed knowledge of his evaluation of positions in
a game. Using this information, and assuming the opponent to be
fallible, he may be outwitted by anticipating his errors.
The risks of doing so are discussed. With respect to possible gains
and losses, it is shown that these may be combined into distinct
strategies, four of which are exhibited and characterized by their
presumed balance among them.
Eduardo Morales: Learning Patterns for Playing Strategies.
A first-order inductive framework, PAL, is described, capable of
learning chess patterns from a combination of two sources, viz.
general-purpose knowledge about chess and simple example descriptions
supplied to it. The King and Rook against King endgame is used
to automatically construct a playing strategy learned by PAL.
A discussion about the present and future state of pattern learning
is presented.
Javier Ros Padilla: Estimating Asymmetry and Selectivity in Chess Programs.
This note represents an attempt to extract information from chess
programs (to the source code of which one has no access) on
asymmetry and selectivity. Measures for these concepts are given
and experimental results presented.
Christian Donninger: A la Recherche du Temps Perdu: 'That was easy'.
This note addresses the concept of using the available time in
chess programs, depending on the state of the game and the nature
of the position. Examples illuminate the author's ideas and
experiences.
Report on:
Games: Planning and Learning.
In this report Jonathan Schaeffer gives a summary of papers presented
at the AAAI Fall Symposium, held at Raleigh, North Carolina, October
22-24, 1993. The papers presented there covered a wide range of topics
applied to a large number of games. Schaeffer gives in the present
report a summary of the papers likely to be of most interest to
computer-chess researchers.
The present issue, containing much more, totals 48 pages of information,
relevant to all computer-chess enthusiasts!!
======================================================================