Author: Christophe Theron
Date: 04:40:52 04/04/02
Go up one level in this thread
On April 03, 2002 at 15:01:22, Chessfun wrote:
>On March 30, 2002 at 16:48:38, Kurt Utzinger wrote:
>
>>A lot of people surprised about the results of Chess Tiger 14 in our tournament
>>argue that there must be something wrong with the settings and so on. That's
>>nonsense, but I am going to check this on the basis of some critical positions.
>>By the way: I am convinced that the hour of ChessTiger14 in this tournament is
>>still to come!!
>
>
>Clearly as has already been written any of the top programs are capable of
>winning a tournament, even from Nunn 1.
>
>As an example I post your first 4 games of Tiger 14 v Fritz 7 006.
>First 2 from position 1, then my 2 from position 1.
>Then your 2 from position 2, and then my 2.
>
>Point swing;
>Your results
>Tiger 14 001=
>
>Mine
>Tiger 14 ==11
>
>A swing of 1-1/2 points in only 4 games.
>
>My games are the same time controls, TB at 1.3 ghz, ponder=off.
>
>Couple of small differences;
>All 3, 4 and 5 piece tablebases and hash Fritz 128 Tiger 96.
>The 64 your using I assume converts to only 48 for Tiger at an average move of 3
>mins I'm not sure that's enough. Maybe Christophe could comment.
Yes, Tiger cannot use 64Mb for the hash tables. It's either 48 or 96, but it
cannot be 64.
On the other hand, using 48 instead of 96 is not going to result in a dramatic
playing strength degradation.
IIRC, doubling the hash size result in a 6% speed increase. That turns into a 6
elo points improvement.
I cannot be an excuse for Tiger's poor result in this tournament, even if it is
clearly not an optimal setup.
What I would really like is to have a look at the CT.INI file of Tiger in this
tournament.
Christophe
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.