Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: "Effectively" Comp GM strength question is answered!

Author: James Robertson

Date: 23:36:41 07/14/00

Go up one level in this thread


On July 15, 2000 at 00:34:08, blass uri wrote:

>On July 14, 2000 at 22:42:33, James Robertson wrote:
>
>>On July 14, 2000 at 19:22:16, Drazen Marovic wrote:
>>
>>>On July 14, 2000 at 19:00:32, Mogens Larsen wrote:
>>>
>>>>On July 14, 2000 at 18:40:10, Drazen Marovic wrote:
>>>>
>>>>>  Scientifically is a comp GM strength?  According to some here no, though i'm
>>>>>not convinced of that oppinion either.
>>>>
>>>>Science has nothing to do with opinion.
>>>
>>>Apparently reading has nothing to do with understanding either!  It is an
>>>oppinion currently as to whether their is enough evidence to say scientifically
>>>if a comp is GM strength.  Further again this term of "GM strength"  Does it
>>>merely mean a statistical result or is it really more substantive to refer to
>>>qaulity of play!  If your average bootom of the barrel 2500 GM played in 3 round
>>>robin 10 game events with Kasparov anand And Kramnik The bottom of the barrel
>>>2500 GM would probably most likely not get his GM Norm much less get the 3
>>>required. That would not necessarily mean that he didn't demonstrate "GM
>>>strength" play.  Geesh the current world champion is barely going to get a GM
>>>norm in Dortmund and Kasparov isn't even there!
>>
>>Nisipeanu's brief flash across the chess scene last year by overcoming many
>>strong players at Las Vegas suggested he was of much greater than >2700 strength
>>(after 10 odd games or so). But that impressive performance turned out to be
>>fluke, and nothing more.
>>
>>The only way to gauge his "true" srength (or close to it) was to look at scores
>>of his games, and they lead to a very different conclusion.
>>
>>Another example: P.Conners in his latest tournament had a TPR higher than what
>>Junior had earned, and after more rounds! Unfortunately for P.Conners, two games
>>dropped it's TPR to less than 2550. Amazing how things can turn around in mere
>>hours....
>>
>>Take another example: poor Luke McShane. After 9 games in Lippstadt, you would
>>think he was the most terrible player in history to play in an international
>>tournament. However, it is better not to write him off, as looking at the "big"
>>picture (dozens and dozens of games by him) tells of high IM strength.
>>
>>Look at the TPR difference between Fritz's performance at the Isreali League
>>(sub 2450!!!!) and at the Dutch Championship! Obviously, the few games played at
>>either event is not enough to say anything susbtantial about Fritz's "true"
>>strength.
>
>I believe that the reason that Fritz failed in the Israeli league is the fact
>that the teams had the right to choose the player to play against the computer.
>
>I believe that we should count only tournaments when everyone has to play
>everyone and not tournament that humans created in order to humiliate the
>computer and not in order to get information about the real strength of the
>computer.
>
>Uri

I disagree. It played those games against those players and had those results.
If the programmer does not like the conditions of the tournament, play in a
different one, as any human with the same problem would.

James



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.