Author: martin fierz
Date: 13:35:56 03/18/04
Go up one level in this thread
On March 18, 2004 at 10:03:33, Tord Romstad wrote: >On March 18, 2004 at 09:34:48, martin fierz wrote: > >>On March 18, 2004 at 05:09:35, Tord Romstad wrote: >> >>>On March 18, 2004 at 00:25:09, martin fierz wrote: >>> >>>>on my machine (P4 1.4GHz) Muse is fastest: >>> >>>... and here, of course (P4 2.4 GHz), Gothmog is fastest. Rxe6 is found >>>after 7 seconds. By the way, I agree that Muse seems to be improving >>>very quickly. >> >>thx but.... >> >>>Gothmog has had a lot of trouble against it on the ICC >>>lately. >> >>...i see 6-2 for gothmogX in march. am i missing something? > >Yes. As so often, the results don't tell the whole story. If you take >a look at the games, you will see that Muse could easily have scored more >than 50% in these games. GothmogX was very lucky. there is no such thing as luck in chess. here in zurich there used to be a bulgarian IM named donev (i think he's changed profession and plays either backgammon or poker these days) who said: "the weak players count the winning positions, the strong players count the wins" :-) if gothmog is outsearching muse by a speed factor of 3-4, then it's rather natural that at some point muse will blow it. or i might have a bad eval term somewhere which you don't etc. >>another thing i should work on is the trojan sac on g5 (or do some learning...), >>some crazy guy on ICC drained muse during the last days by repeating the same >>game whenever he got a chance. > >I think a good king safety eval is a better way to handle this than adding >special code for trojan sacs. i don't do x-ray attacks (yet). so muse doesn't understand that Rh1+Qh5 is dangerous (it would understand that Ng5 or Bd3 and Qh5 is dangerous). that is rather obviously a bit of a drawback :-) i fixed it now by inventing a new search extension which triggers in this case. however, it's something general, not a pattern-match to the trojan (like i think crafty does) - i don't like the idea of doing "if X then Y" and then writing down zillions of Xs... cheers martin
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.