Author: Michael Neish
Date: 07:15:14 01/07/06
Go up one level in this thread
On January 07, 2006 at 10:00:22, Alessandro Scotti wrote: >On January 07, 2006 at 09:26:16, Michael Neish wrote: > >>With such a high rate, maybe the benefit is more obvious. Out of curiosity, >>have you any idea how much more depth null move is giving you? Is it enough to >>add a whole play, say from 9 to 10? > >I run the usual 5 seconds test on a subset of WAC (posted below). Machine is a >700 MHz Athlon. > >Kiwi scores 10/13 with null move, and 6/13 without. Null move seems to give 1 or >2 extra plies. Here's a detailed report: > > 8/13 vs 7/13 > 9/14 vs 7/11 >12/36 vs 13/38 (yes, got deeper without NM on this one) > 8/11 vs 8/13 > 9/15 vs 7/17 > 8/13 vs 7/12 >10/16 vs 8/12 >12/19 vs 9/13 > 8/15 vs 8/15 >12/18 vs 10/16 > 9/12 vs 8/16 > 8/17 vs 8/13 > 9/13 vs 8/13 > >2kr3r/pp1q1ppp/5n2/1Nb5/2Pp1B2/7Q/P4PPP/1R3RK1 w - - bm Nxa7+; id "WAC.071"; >r4rk1/1p2ppbp/p2pbnp1/q7/3BPPP1/2N2B2/PPP4P/R2Q1RK1 b - - bm Bxg4; id "WAC.092"; >8/k1b5/P4p2/1Pp2p1p/K1P2P1P/8/3B4/8 w - - bm Be3 b6+; id "WAC.100"; >4r1k1/p1qr1p2/2pb1Bp1/1p5p/3P1n1R/1B3P2/PP3PK1/2Q4R w - - bm Qxf4; id "WAC.141"; >5rk1/2p4p/2p4r/3P4/4p1b1/1Q2NqPp/PP3P1K/R4R2 b - - bm Qg2+; id "WAC.163"; >r1q2rk1/p3bppb/3p1n1p/2nPp3/1p2P1P1/6NP/PP2QPB1/R1BNK2R b KQ - bm Nxd5; id >"WAC.180"; >1r2k1r1/5p2/b3p3/1p2b1B1/3p3P/3B4/PP2KP2/2R3R1 w - - bm Bf6; id "WAC.183"; >r3k3/P5bp/2N1bp2/4p3/2p5/6NP/1PP2PP1/3R2K1 w q - bm Rd8+; id "WAC.221"; >2r1r2k/1q3ppp/p2Rp3/2p1P3/6QB/p3P3/bP3PPP/3R2K1 w - - bm Bf6; id "WAC.222"; >2b5/1r6/2kBp1p1/p2pP1P1/2pP4/1pP3K1/1R3P2/8 b - - bm Rb4; id "WAC.230"; >1r3r1k/3p4/1p1Nn1R1/4Pp1q/pP3P1p/P7/5Q1P/6RK w - - bm Qe2; id "WAC.243"; >rn3k1r/pp2bBpp/2p2n2/q5N1/3P4/1P6/P1P3PP/R1BQ1RK1 w - - bm Qg4 Qh5; id >"WAC.287"; >3r1rk1/p3qp1p/2bb2p1/2p5/3P4/1P6/PBQN1PPP/2R2RK1 b - - bm Bxg2 Bxh2+; id >"WAC.297"; Wow, thanks for that. I expected a quick reply -- I didn't think you'd go and actually run some tests. That looks pretty conclusive, I guess, so I ought to take a long, hard look at how I'm implementing it. That third position looks to be one of those borderline ones where null move doesn't work very well. I wonder how many programs would turn off null move with that amount of material left. Cheers, Mike.
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.