Author: Albert Silver
Date: 03:56:23 11/16/04
Go up one level in this thread
On November 16, 2004 at 01:46:47, Kurt Utzinger wrote:
>On November 15, 2004 at 15:42:01, Albert Silver wrote:
>
>>On November 15, 2004 at 13:37:01, Kurt Utzinger wrote:
>>
>>>Nunn2 match (60m+10s) 40 games
>>>Pro Deo 1.0 vs Fritz 8.0.0.26
>>>http://www.utzingerk.com/pro_deo.htm
>>>
>>>A little test should show the influence of the option [clear hash table =
>>>always] and [clear hash table = never]. Below the result against Fritz 8 as Pro
>>>Deo had the worst result against this engine:
>>>
>>>Option [Clear Hash Table = Always]
>>>PD = 32.5 %
>>>13.0 - 27.0
>>>
>>>[Clear Hash Table = Never]
>>>PD = 36,2 %
>>>14.5 - 25.5
>>
>>I ran some Nunn2 blitz matches as you know, with and without, and the results
>>were not clear. Against Aristarch and Ruffian the results were better, and
>>against List, PD scored 1 point less.
>>
>> Albert
>
>
> Hi Albert
> Is it worth to replay all remaining 280 games
> at time control 60m+10s with clear hash table
> = never http://www.utzingerk.com/pro_deo.htm
> Kurt
I can't really answer that question as it is your computer time, and that isn't
asking a little amount. It *would* permanently answer the question whether it
plays better overall with one or the other in matches, but I have mixed feelings
about this. As mentioned already, with it set to Never, there are unfortunate
moments where it will choose 2nd, 3rd, or 4th best moves in its list, sometimes
leading to disastrous results (unexpected loss/draw of the game). I have never
heard of such a thing with other top engines, so presume this means there is a
problem/bug, yet Ed is adamant in saying it is not.
What if the bug is small enough to not completely kill the score so that with
Never it still scores 2% better (or whatnot). It might be incorrectly concluded
that there's no bug and that it rightly wins more. However, what if the expected
margin should be 8% better (I'm just inventing numbers here BTW)? The only way
to really know, would be to test another engine, presumed with no bug, with and
without clearing hashtables and see how big a difference is seen in the results,
and compare with PD to see if it shows a similar difference. If there were a
visible difference, I think Ed might be motivated to try and hunt the problem
done, other wise he may think I'm making too much out of nothing. I am NOT
suggesting you do this BTW, as this would be very time consuming. It just
represents my thoughts as I do not think he is convinced there is a problem, or
that it is worth tracking down.
I will say that I play all my test matches with the setting at Always, merely
because I don't only want reproducible results/moves, but I also do not want to
worry whether a move that I see played is because of my settings or some bizarre
hashtable issue. I'll be honest with you in saying that I have seen moves played
by Pro Deo, even with Always, that I cannot reproduce ever. Meaning the move is
NEVER its favorite choice at any moment.
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.