Author: Ed Schröder
Date: 02:07:34 01/15/01
Go up one level in this thread
On January 13, 2001 at 21:44:45, Howard Exner wrote: >Here are some results by Tiger 13 and Century 3on an Athlon 900. > >Tiger 13 avoids Qe4 at 3:03 by choosing Ka6 at depth 13. >Later on the same ply it changes to b3 at 4:27 then b4 at 7:21. > >Century 3 has this output > >0:52 12th ply Qe4 +1.66 >2:18 13th ply Qe4 -1.29 > >Will not avoid the move Qe4 till the next >iteration in 14:22, playing b4. > >That seems like quite a long delay as the eval swung +2.95. >Given that Qe4 had -1.29 after 2:18 when it took an extra 12:04 >to resolve that b4 is better, why such a long delay? It has something to do with move-ordering. In some extreme cases of a fail-low (drop in score) many critical best-moves from the hash table simply fail and thus other moves must be tried till the best one is found. In general: if the search finds the best move (on a given depth) soon there will be not so much delay but if the search guesses wrong time after the time huge time delays are the result. Ed PS, while writing I tried C2 under the Tiger GUI, it also takes a long time to resolve the fail-low (0:58 <> 5:12_ Loading... Engine version : Rebel Century 2.01 Hash table size : 40 Mb 5k2/p4pp1/1p5p/1K2P3/5q2/P4P2/1P2Q1P1/8 w - - 00:00 04.00 0.61 1.Qe4 Qxe4 2.fxe4 Ke7 3.Ka6 b5 4.Kxa7 Ke6 5.Kb6 Kxe5 00:01 06.00 0.61 1.Qe4 Qxe4 2.fxe4 Ke7 3.Ka6 b5 4.Kxa7 Ke6 5.Kb6 Kxe5 6.Kxb5 Kxe4 00:01 07.00 0.61 1.Qe4 Qxe4 2.fxe4 Ke7 3.Ka6 b5 4.Kxa7 Ke6 5.Kb6 Kxe5 6.Kxb5 Kxe4 00:02 08.00 1.18 1.Qe4 Qg5 2.Qa8+ Ke7 3.Qxa7+ Ke8 4.Qxb6+ Qxg2 5.Qc6 Ke7 00:04 09.00 1.43 1.Qe4 Qg5 2.Qa8+ Ke7 3.Qxa7+ Ke8 4.Qb8+ Ke7 5.Qd6+ Ke8 6.g4+ h5 7.Qc6+ Ke7 00:08 10.00 1.25 1.Qe4 Qg5 2.Qa8+ Ke7 3.Qxa7+ Ke8 4.Qb8+ Kd7 5.g3 h5 00:21 11.00 1.40 1.Qe4 Qg5 2.Qa8+ Ke7 3.Qxa7+ Ke8 4.Qb8+ Kd7 5.Qd6+ Ke8 6.Kc6+ h5 7.a4 00:58 12.00 -0.47 1.Qe4 Qxe4 2.fxe4 g5 3.Kc4 h5 4.a4 Ke7 5.b4 Ke6 6.Kd4 h4 7.a5 b5 8.a6 05:12 12.01 0.00 1.Qe1 Qh2 2.Qe4 Qxg2 3.Qa8 Ke7 4.Qb7 Ke6 5.Qe4 Qxb2+ 6.Kc6 Qxe5 7.Qc4
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.