Author: Uri Blass
Date: 10:32:11 04/04/03
Go up one level in this thread
On April 04, 2003 at 12:47:31, Robert Hyatt wrote: >On April 04, 2003 at 10:43:54, Uri Blass wrote: > >>On April 04, 2003 at 10:13:19, Jorge Pichard wrote: >> >>>On April 04, 2003 at 09:10:22, Uri Blass wrote: >>> >>>>Fritz8 lost on time 2 games(one in a drawn position and one in a won position) >>>>but time is part of the game. >>>> >>>>games were played on A1000(16 mbytes hash tables for every engine) >>>> >>>>Uri >>> >>>I wonder how bad Fritz 8 will beat List 5.04 if you allow 2 minutes to Fritz 8 >>>instead? >>> >>>Jorge >> >>I do not know >>You can try to play a match of 2 minutes for Fritz against 3 minutes for list. >> >>I guess that it may beat List 12-8 in the nunnI match. >> >>I also wonder if Fritz performs better in the nunnI match or in my special match >>(40 games from all the possible positions after 1 ply). >> >>I saw a lot of cases when Fritz disagreed with the opponent about evaluation. >>It happened also in the match against Tao5.4 when Fritz8 beated Tao5.4 15-5 >> >>For some reason my feeling is that the nunnI match give relatively more >>positions when programs disagree about the evaluation(difference of more than 1 >>pawn between the evaluation of the programs). >> >>Fritz8 seems to be a positional monster at least in the nunnI match >> >>examples >> >>[D]3r3k/4q2p/5p2/3p1Q2/R7/3Bb1P1/7P/7K b - - 0 39 > >This is not position. This is tactics. Crafty comes up with 0.00 very quickly, >but >not positionally. It is a repetition. > > >> >>Fritz 0.00 list advantage of more than +1 for black >> >>[D]8/5k1p/8/8/P1R5/3pp1bP/8/3K4 w - - 0 49 >> >>Fritz >> >>more than +1 for white > >Tactics again. Takes 16 plies for Crafty to see past a perpetual and fail high >on >Re4 to +1.0 > > >>List:position almost equal >> >>[D]r2qb2k/1n3pp1/r6p/p2NP3/2pPB3/1pP2Q2/1BP3PP/5RK1 b - - 0 33 >> >>Fritz more than +1 for white > >This is also tactics. The other two crafty found very quickly. This takes >about 45 seconds >before it sees black is in trouble, at depth=12, score = +1. My opinion is that Fritz simply suspected the tactics for positional reasons. In theory every positional knowledge is tactics if you search deep enough. I agree that there are cases when the tactics is so deep that practically search is not going to find it. I also agree that in the relevant cases it is also possible to see things for tactical reasons but I do not believe that Fritz saw it for tactical reasons in 1 minute for game on A1000. I did not install Fritz8 on the computer that I use now so I will be able to check fritz8's evaluation at small depthes only later. Uri
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.