Author: John Roberts
Date: 06:22:09 01/15/04
Go up one level in this thread
On January 15, 2004 at 09:08:52, George Tsavdaris wrote: >On January 15, 2004 at 08:52:00, John Roberts wrote: > >>A - B >> >>6k1/5p1p/1nB1n1p1/1P6/2pq2P1/3p3P/1P1Q1PK1/R7 w - - 0 1 >> >>Analysis by Fritz 8: >> >>>37.Qc3 Qd6 38.Re1 Na4 39.Qd2 Kg7 40.Ra1 Nb6 41.Qc3+ Kg8 >> ² (0.31) Depth: 19/55 17:27:52 32245654kN >>37.Qc3 Qd6 38.Re1 Na4 39.Qd2 Kg7 40.Ra1 Nb6 41.Kf1 Nd4 >> ² (0.31) Depth: 20/56 45:57:00 87323100kN >>37.Qc3 Qd6 38.Re1 Na4 39.Qd2 Kg7 40.Ra1 Nb6 41.Kf1 Nd4 >> ² (0.28) Depth: 21/74 121:11:44 234690340kN, tb=1 >>37.Qc3 Qd6 38.Re1 Na4 39.Qd2 Kg7 40.Ra1 Nb6 41.Kf1 Nd4 >> ² (0.28) Depth: 21/88 318:20:01 625183936kN, tb=16 >> >> >>The clock readings are correct, so why doesn't the last line read 22/88? >> >>Pentium4 2.4GHz with 334MB hash, 512MB ram compaq presario. > >Wow! 318 hours. Why do you care so much about this position? > >I have seen Fritz do this, thousants of times. It shows a move in depth >X/Y and it shows it again at depth X/Y+K (K>0). Why do you think it's strange? >Look for example the following at depth 9 and 11. The same happens: > >Analysis by Deep Fritz 8: > >46...Qf4xd6 47.a2-a4 Qd6-d4 48.d7-d8Q Qd4xd8 49.f2-f4 Qd8-d4 50.Bb5-c4 > -+ (-3.57) Depth: 7/21 00:00:00 50kN >46...Qf4xd6 47.a2-a4 Qd6-d4 48.d7-d8Q Qd4xd8 49.f2-f4 Qd8-d4 50.Bb5-c4 Qd4-b2+ > -+ (-3.57) Depth: 8/21 00:00:00 81kN >46...Qf4xd6! > -+ (-3.86) Depth: 9/25 00:00:00 192kN >46...Qf4xd6 47.a2-a4 Qd6-d4 48.d7-d8Q > -+ (-4.11) Depth: 9/29 00:00:00 347kN >46...Qf4xd6 47.Bb5-c4 Qd6xd7 48.f2-f3 Qd7-d4 49.f3xe4 Qd4xe4+ 50.Ke2-d2 Qe4-g2+ >51.Kd2-e3 Qg2xh3+ 52.Ke3-d4 Bh4-f6+ 53.Kd4-c5 Qh3-e3+ 54.Kc5-b5 Kh7-g6 > -+ (-4.34) Depth: 10/27 00:00:01 987kN >46...Qf4xd6! > -+ (-4.62) Depth: 11/29 00:00:03 1804kN >46...Qf4xd6! > -+ (-4.90) Depth: 11/32 00:00:05 2444kN >46...Qf4xd6 47.Bb5-c4 Qd6xd7 48.f2-f3 Qd7-d4 49.f3xe4 Qd4xe4+ 50.Ke2-d1 Qe4-b1+ >51.Kd1-d2 Qb1xa2+ 52.Kd2-c3 Qa2-a5+ 53.Kc3-c2 Bh4-g5 54.Rf1-d1 Bg5-f6 55.Bc4xf7 > -+ (-5.09) Depth: 11/32 00:00:07 3472kN Of course this happens, but in your examples, it doesn't circulate through all the possible candidates before spitting out the 2nd (or 3rd, etc) variation with the same key move. The example I gave, I've never seen anything like it, as part of the 197 hours were spent examining each candidate once. The only buggy aspect would be that I've watched DVDs and surfed while this (infinite) analysis is currently performed, with tablebases enabled. A friend asked me to help figure out if the position is won for White, intellectual curiousity is all. See posts about 2 months ago...
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.