Xref: utzoo rec.games.go:972 comp.ai:4010 rec.games.board:2083
Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!cs.utexas.edu!uunet!mcvax!jansteen
From: jansteen@cwi.nl (Jan van der Steen)
Newsgroups: rec.games.go,comp.ai,rec.games.board
Subject: Re: Computer Go Challenge
Summary: Relation between strength and result
Keywords: go, computer games, ai
Message-ID: <520@turing.cwi.nl>
Date: 26 Apr 89 19:13:34 GMT
References: <3724@sdsu.UUCP>
Organization: CWI, Amsterdam
Lines: 49



In the Dutch Go magazine 16-6 (august 1979) an article was
published which amongst other dealing with the classification
of Go players. The author of this article is Prof. Dr. K. Heine
(Wilhelmshaven, Germany).

= Quote start =

In the next graph, different results of Go games is measured against
the strength of the players involved:


    spread in results
            |
          50|                                             .
            |                                        .
          40|                                   .
            |                              .     *
          30|                         *
            |                    .  *
          20|               *
            |          *  *
          10|     .*   
            |. *
       _.___|_______________________________ strength (Japanese system)
            |     |     |     |     |     |
            0     5D    1k    5k    10k   15k


        Some conclusions:

        - Perfect Go is played by someone who is able to give a
          9 Dan professional a four stone handicap.
        - The average size of the result is proportional to the
          quality of the game.
        - ...

= Quote end =


Interesting, isn't it?


-- 
	-=-=-=-=-=-=-=-=-=-=-=-=-=--=-=-=-=-=-=-=-=-=-=-=-=-
   	 Jan van der Steen,		CWI the Netherlands
   	 jansteen@cwi.nl	  (or uunet!mcvax!jansteen)
	-=-=-=-=-=-=-=-=-=-=-=-=-=--=-=-=-=-=-=-=-=-=-=-=-=-