Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Alpha/Beta adjustments via the hashtable or why Fritz 7 has messy PV's

Author: Tony Werten

Date: 04:21:06 12/04/01

Go up one level in this thread


On December 04, 2001 at 06:27:41, Ulrich Tuerke wrote:

>On December 04, 2001 at 03:55:21, Tony Werten wrote:
>
>>On December 03, 2001 at 11:01:56, Ulrich Tuerke wrote:
>>
>>>On December 03, 2001 at 09:26:21, Gian-Carlo Pascutto wrote:
>>>
>>>>On December 03, 2001 at 09:19:16, Ulrich Tuerke wrote:
>>>>
>>>>>I don't agree fully with you, Gian-Carlo.
>>>>>
>>>>>I think that getting unreliable main variations is a well known consequence of
>>>>>any zero window search like for instance PVS or MTD.
>>>>>The reason is that you will very often have moves in the PV which are just
>>>>>sufficient to refute the preceeding move. In general you wan't have exact scores
>>>>>neither exact variations.
>>>>>
>>>>>I don't think that this effect has to be attributed to the hashing algorithm.
>>>>
>>>>In PVS, the mainvariation should _not_ have been searched with
>>>>a zero window, and should hence be reliable.
>>>
>>>I think that you're right. My remarks apply to MTD only.
>>
>>It is actually the same reason why it's so difficult to get a good pv when using
>>MTD.
>
>
>So, may be Fritz has changed to MTD ?
>We'll ask Frans or Mathias in Paderborn.
>
>Uli
>
>BTW, don't play soccer these days.

???

Tony

>>
>>Tony
>>
>>>Uli
>>>>
>>>>--
>>>>GCP



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.