Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Fail-soft or Fail-hard ?

Author: Robert Hyatt

Date: 06:38:02 11/04/00

Go up one level in this thread


On November 04, 2000 at 07:42:00, Ulrich Tuerke wrote:

>On November 04, 2000 at 00:48:52, TEERAPONG TOVIRAT wrote:
>
>>Hi,
>>Is fail-soft alpha beta too slow?
>>I see most programs prefer fail-hard version.
>>Thanks for any comment.
>>Teerapong
>
>IMO, either methods has its benefit. I understand that fail-hard is a bit faster
>than fail-soft because "best" is already initialized to alpha, whereas it's "-
>infinity" in fail-soft.
>
>So in case, the search does successfully return a value out of the initial
>alpha-beta-window, I expect fail-hard to be a bit less node-consuming.
>
>However, if you encounter a value outside the initial alpha-beta region, the
>fail-soft algo may give you a bound out side alpha-beta, whereas fail-hard can
>at most give the bounds alpha or beta itaself.
>
>Uli


I'm not aware of anything that says that when using fail-soft, you _must_ start
off at +/- infinity.  I don't, for example, yet I still let the search return
values outside alpha/beta.  I found it slightly better because you get a bit
better info out of the hash table.  But it was only a "bit" better.  Not a huge
win.

More important if you use mtd(f) of course.



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.