Newsgroups: comp.ai.games
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!gatech!howland.reston.ans.net!news.nic.surfnet.nl!sun4nl!aie.nl!news
From: geert@sparc.aie.nl (Geert-Jan van Opdorp)
Subject: Re: Quality measure for games
Sender: news@aie.nl (News Account)
Organization: AI Engineering BV
Lines: 42
Message-ID: <GEERT.95Apr27183953@sparc.aie.nl>
References: <95117.143021HLR002@DJUKFA11.BITNET>
In-Reply-To: HLR002@DJUKFA11.BITNET's message of 27 Apr 95 12:28:21 GMT
Date: Thu, 27 Apr 1995 16:39:53 GMT


Hi,

>>>>> "Manfred" == Manfred Kremer <HLR002@DJUKFA11.BITNET> writes:

Manfred> What I would like to know is this: Without looking at the
Manfred> game tree and checking each player's moves against the
Manfred> optimal moves: is there any numerical measure to compare a
Manfred> good game to a poor game? Intuitively I feel there should be
Manfred> a way, as we can distinguish beteen a high call game and a
Manfred> beginners's game, but has this ever been formalized.  I am
Manfred> specifically interested in this question related to go.

In go there may be a number of characteristics that reflect
the quality of the game. For example it might
be interesting to make a statistical analysis of the number
of tenuki's (moves that are not tacticly related to the
immediately preceding moves) played in games by players 
of different ranks.
Only, ofcourse, it is hard to give a formal definition of 
`tenuki'. 

Another indication of the quality of a collection
of games (or maybe rather of the strength of the players)
may be the variation in the margin by which the game is won.
There seems to be a very clear relation of this variation
and the strength of the players. There was an article written
about this, maybe someone on rec.games.go can point you
to it. 
Interrestingly, this relation was used to estimate the
strength of optimal play. This was done by extrapolating
the curve to the point where there would be no variation in
results. According to this, `God' would be about 2 or 3 stones
stronger than the top professionals. This is like the
gap between the weakest and the strongest professionals.

Geert-Jan
-- 
Geert-Jan van Opdorp
AI-Engineering
Amsterdam, The Netherlands
geert@aie.nl
