Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: 1 minute Blitz---------very interesting

Author: Uri Blass

Date: 07:30:59 09/27/03

Go up one level in this thread


On September 27, 2003 at 09:49:45, Russell Reagan wrote:

>On September 27, 2003 at 04:43:27, swaminathan natarajan wrote:
>
>>then which is the best time control for blitz games?
>>taht will show the engines original strenght
>
>I don't think there is a "best" time control. If you want to know the "best"
>engine, probably longer time controls are better. If you want to know the
>program with the fewest and least significant weaknesses, probably very fast
>games will show that.
>
>Take SmarThink or Yace for example. These engines, in my experience, have bad
>results against a program like Ruffian in 1 minute games. I think the reason is
>not that they are THAT much weaker than Ruffian, but that their time management
>probably isn't as sophisticated, and in a 1 minute game that can cost you the
>game. So SmarThink and Yace have a weakness in their time management. At longer
>time controls, this weakness is covered up.

Time management is not the strong point of ruffian.
Try it at 10000 moves/100 minutes and you are going to find that it does not use
its time.

I think thst Ruffian is clearly better than Yace at long time control and the
result of Leo tournament proves it.

Uri
>
>The same thing happens for all kinds of weaknesses. If a program lacks some kind
>of knowledge, it can cover up that weakness by using extra search. In a 1 minute
>game, that weakness will be exposed.

I am not so sure.
There are weakness that are not important in 1 minute/game but are important at
longer time control.

 This is why you often hear programmers
>saying that their engine "performs better at longer time controls", or how
>"<insert engine here> is very strong at longer time controls."

I do not hear often programmers say it.

>
>It isn't really stronger. Rather, fewer of its weaknesses are being exposed.

also fewer weaknesses of the opponent are exposed so it is not a reason to
perform better at longer time control.

Uri



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.