Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Mercilous Attack? What is it? What is the ECO on it?

Author: Bruce Moreland

Date: 18:42:35 07/07/99

Go up one level in this thread



On July 07, 1999 at 20:05:51, Dann Corbit wrote:

>On July 07, 1999 at 18:06:05, Robert Hyatt wrote:
>[snip]
>>You missed an _important_ detail.  This program was running on slower hardware,
>>and the game time control was game in 3 minutes, no increment.  This makes it
>>easy to overlook the mate...  This _only_ worked at very fast time controls,
>>and only against slower hardware.  But once the human figured it out, it was
>>difficult for _any_ program that could play that fast to handle.  I stopped
>>losing those games once I got to faster hardware, but left the eval tweak in
>>for those that run into it...
>It seems that there is a hidden sucker punch that people can tap into.  If you
>can estimate the depth in plies that a computer will play at for a given time
>control and machine setup, then you can set a trap one ply beyond that horizon.
>Or is there some other ingredient that makes the trick work?

You'd think that if someone had enough time to consider and verify something
like that, that they might also be able to simply play strong moves and beat it,
and that this technique would generalize better, meaning that they could use it
against humans as well.

The Mercilous Attack worked because it played on a computer weakness, the
inability to correctly evaluate attacking chances.

bruce



This page took 0.02 seconds to execute

Last modified: Thu, 07 Jul 11 08:48:38 -0700

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