COMMENT ⊗   VALID 00003 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	Proposals
C00003 00003
C00005 ENDMK
C⊗;
Proposals

One class tournaments.

Inverse killer.

set of standard test positions - comparing the amounts of lookahead
of different algorithms on the standard set.

higher level concepts



In the 1960s the late Soviet computer
scientist Alexander Kronrod referred to chess as the Drosophila
of artificial intelligence.  He hoped that it would play a role
in AI similar to that of the fruit fly in genetics --- an
experimental vehicle for discovering results of general interest
in the subject.  I think Kronrod was right in hoping for this,
but the results have often been disappointing.

	Computer chess became more of a sport than a vehicle
for science.  Too much of it has been characterized by a search
for the fastest possible hardware and programming tricks.  Most
of the programs have not resulted in scientific articles.  The
games played by the program are too often analyzed in the same
way as human games are analyzed, i.e. no advantage is taken of
the fact that it is possible to make the computer run through its
move tree and tell us whether certain positions came up on the
tree and what values they were given.

The lecture will discuss some proposals for making computer chess
more useful to the science of AI.