Author: Albert Silver
Date: 08:17:36 09/02/04
Go up one level in this thread
On September 02, 2004 at 02:49:00, Ed Schröder wrote:
>On September 01, 2004 at 21:05:24, Albert Silver wrote:
>
>>On September 01, 2004 at 20:20:59, Ed Schröder wrote:
>>
>>>On September 01, 2004 at 19:39:29, Albert Silver wrote:
>>>
>>>>On September 01, 2004 at 18:59:42, Ed Schröder wrote:
>>>>
>>>>>On September 01, 2004 at 17:16:09, Kurt Utzinger wrote:
>>>>>
>>>>>>Hello Ed
>>>>>>Is this also your recommendation for my test suite at
>>>>>>[http://www.utzingerk.com/rebel___pro_deo.htm] or should
>>>>>>I continue with default settings?
>>>>>>Kind regards
>>>>>>Kurt
>>>>>
>>>>>I would recommend the personality I just posted.
>>>>>
>>>>>[My Queen = 101] * new since September 1, 2004
>>>>>[Pruning = MISC_78] *
>>>>>
>>>>>These 2 new stettings were not sufficiently tested at release time but in the
>>>>>meantime they fully are and add +10-15 elo.
>>>>>
>>>>>My best,
>>>>>
>>>>>Ed
>>>
>>>
>>>>Another necessary one when playing comp-comp matches through the Fritz GUI is:
>>>>
>>>>[Clear Hash Table = Always] * Always|Never|PB only
>>>>
>>>
>>>
>>>I tested this because of the noise it recently gave, it's nonsense.
>>>
>>>My best,
>>>
>>>Ed
>
>
>>I´m not sure I understand. Are you suggesting that it isn´t necessary?
>
>Yes, just can safely use the default setting, actually I recommend it.
>
>
>>I could always send the games and log, but you yourself were able to
>>reproduce the problem per your words.
>
>I already gave the explanation for the occasional move differences, that is, if
>you change the move-ordering of a (highly) selective program (all programs are
>these days) you will get different moves so now then, it comes as an unwished
>bonus, no way around it. Clearing versus not clearing the hash table does just
>that, it changes the move-ordering.
>
>I played some volume eng-eng match NEVER vs ALWAYS and NEVER was slightly
>victorious, just as expected.
>
>Hope this helps,
>
>Ed
Yes, it does help, and thanks for looking into detail. I am still really irked
that at 14 plies deep it means it can miss a move that loses in 4 plies as it
did in the example, but I guess the total gains outweigh the losses.
Albert
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.