Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Thanks for telling me its strength is not positional!

Author: Albert Silver

Date: 19:31:12 01/14/06

Go up one level in this thread


On January 14, 2006 at 22:11:01, Albert Silver wrote:

>On January 14, 2006 at 21:50:14, stuart taylor wrote:
>
>>I had been under the impression that Rybka's main strength was positional, and
>>that others came no where near it. I thought I had heard it was tactically good
>>too.
>>But thanks for letting me know it was purely tactical. I had been under the
>>impression for some reason that the ice was beginning to be broken from the
>>stagnation of some of chess computing weaknesses.
>>So there is still way to go, and Rybka might yet be succeeded by about another
>>200 elo, by positional improvement alone!
>>S.Taylor
>
>He's wrong.
>
>      Albert

I should add that the evidence he is wrong is absolutely *huge*. He is basing
his statement on an analysis by Lacrosse showing Rybka is a tactical beast. Of
course, there is no question of that. Without tactics, it couldn't possibly have
a chance.

Lacrosse's analysis showed above all that in the 87 positions he tested, that
Shredder 9 and Rybka scored 57% given 10 seconds, and Fruit and Toga and company
are much weaker with so little time, and thus much weaker in blitz. Well, if
that is true then Shredder 9 should be far ahead of Fruit and Co. on the blitz
rating lists, but of course they aren't. Not on a single one in fact. Let's take
the ever popular CEGT Blitz rating lists
(http://www.husvankempen.de/nunn/eloblitzall.html):

no  	Program  	        Elo  	+  	-  	Games  	Score

4	Rybka 1.0 Beta 32-bit	2824	12	17	1694	69.1 %
5	Rybka 1.0 Beta 32-bit VT2786	16	20	1131	66.8 %
6	Hiarcs 10 Hyperm	2783	22	22	720	60.0 %
7	Fritz 9	                2773	14	17	1551	65.0 %
8	Hiarcs 10	        2768	19	16	1133	55.8 %
9	Fruit 2.2.1	        2761	10	10	3254	61.0 %
10	Fruit 2.2 ht50 ks104	2756	22	28	584	67.3 %
11	Toga II 1.1	        2754	14	14	1655	59.9 %
12	Shredder 9.1 UCI	2739	14	15	1803	60.7 %

As you can see, not only are Fruit and Toga better blitz players (which is no
news actually), but the more tactical setting of Rybka (VT = Very Tactical)
scored considerably worse than the default Very Positional setting.

Mind you, this is just a comment on the numbers. As a player, the games speak
for themselves. I don't mean who wins and by how much, but how the games are
win. It is completely obvious that Rybka is outplaying its opponents (just see
how often it says +0.60 and the other thinks IT is better by a similar score),
and not simply by better tactics.

                                       Albert



This page took 0.01 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.