Author: Shaley
Date: 07:48:54 08/30/05
Go up one level in this thread
On August 30, 2005 at 10:38:13, Uri Blass wrote: >On August 30, 2005 at 10:20:56, Shaley wrote: > >>There is an interesting article on whether and how to use chess engines in >>correspondence games placed in ICCF web zine at >>http://amici.iccf.com/issues/issue_05/issue_05_Chess_engines.html > >I can say only that the author is not a programmer and has no idea how chess >engines work. > >his claim that they look only at part of the moves is clearly wrong. > >1)Even with no unsound pruning engines do not need to look at all possibilities >in order to get depth 12 and his 30^12 number is clearly wrong. > >It seems that the author does not know what is alphabeta and the same result can >be achieved with branching factor of 6 without if you use simple alphabeta with >no pruning. > >2)His claim that they look only at 4 out of 30 moves is clearly wrong. >They may reduce depth for part of the moves but they look at all moves. > >3)In the relevant game of peter leko against Kramnik the mistake of kramnik can >be easily identified by top programs. > >Here is anlaysis of fruit2.1 > >New game - Fruit 2.1, 38'/40+38'/40+38'/40 >[D]6k1/5p1p/p1pb1nq1/1P4p1/3P4/1BPb1PP1/1P1NrQ1P/R1B3K1 w - - 0 1 > >Analysis by Fruit 2.1: > >24.bxa6 Rxf2 25.Kxf2 > -+ (-2.39) Depth: 1/3 00:00:00 >24.bxa6 Bxa6 25.Qf1 Qh6 > ² (0.40) Depth: 1/3 00:00:00 >24.bxa6 Bxa6 25.Qf1 Qh6 > ² (0.40) Depth: 4/13 00:00:00 >24.bxa6 Bxa6 25.Qf1 Qh6 > ² (0.40) Depth: 4/13 00:00:00 >24.bxa6 Bxa6 25.Qf1 Qh6 > ² (0.40) Depth: 4/13 00:00:00 >24.bxa6 Bxa6 25.Qf1 Qh6 > ² (0.40) Depth: 4/13 00:00:00 >24.bxa6 Bxa6 25.Qf1 Qh6 > ² (0.40) Depth: 4/13 00:00:00 >24.bxa6 Bxa6 25.Qf1 Qh6 > ² (0.40) Depth: 4/13 00:00:00 >24.bxa6 Bxa6 25.Qf1 Qh6 > ² (0.40) Depth: 4/13 00:00:00 >24.bxa6 Bxa6 25.Qf1 Qh6 26.h4 > ± (0.80) Depth: 5/18 00:00:00 >24.bxa6 Bxa6 25.Qf1 Qh5 26.h4 Bd3 27.Ra8+ Ne8 > ± (1.05) Depth: 6/18 00:00:00 52kN >24.bxa6 Rxf2 25.a7 Re2 26.a8Q+ Re8 27.Qxc6 Re1+ 28.Kf2 Re2+ 29.Kg1 Re1+ > = (0.00) Depth: 7/23 00:00:00 134kN >24.bxa6 Rxf2 25.a7 Re2 26.a8Q+ Re8 27.Qxc6 Re1+ 28.Kf2 Re2+ 29.Kg1 Re1+ > = (0.00) Depth: 8/25 00:00:00 240kN >24.bxa6 Rxf2 25.a7 Re2 26.a8Q+ Re8 27.Qxc6 Re1+ 28.Kf2 Re2+ 29.Kg1 Re1+ > = (0.00) Depth: 9/29 00:00:00 465kN >24.Qxe2 Bxe2 25.bxa6 Ne8 26.a7 Nc7 27.Ne4 h6 28.Nxd6 Qxd6 29.a8Q+ Nxa8 30.Rxa8+ >Kg7 31.Ra7 > ± (0.88) Depth: 9/29 00:00:01 966kN >24.Qxe2 Bxe2 25.bxa6 Bb8 26.a7 Bxa7 27.Rxa7 Nd5 28.Kf2 Bd3 29.Ne4 h6 30.Bxd5 >cxd5 > ² (0.61) Depth: 10/31 00:00:02 1480kN >24.Qxe2 Bxe2 25.bxa6 Bb8 26.a7 Bxa7 27.Rxa7 Nd5 28.c4 Ne3 29.c5 Qd3 30.Bxf7+ Kf8 >31.Rd7 > ± (0.86) Depth: 11/31 00:00:03 2427kN >24.Qxe2 Bxe2 25.bxa6 Bb8 26.a7 Bxa7 27.Rxa7 Nd5 28.c4 Ne3 29.c5 Qd3 30.Bxf7+ Kf8 >31.Kf2 g4 > ² (0.58) Depth: 12/36 00:00:04 4292kN >24.Qxe2 Bxe2 25.bxa6 Bb8 26.a7 Bxa7 27.Rxa7 Nd5 28.Ra8+ Kg7 29.Re8 Bd3 30.Bxd5 >cxd5 31.Re5 f6 32.Rxd5 Qe8 33.Nb3 Qe1+ 34.Kg2 > ² (0.61) Depth: 13/50 00:00:09 9038kN >24.Qxe2 Bxe2 25.bxa6 Bb8 26.a7 Bxa7 27.Rxa7 Nd5 28.Ra8+ Kg7 29.Re8 Bd3 30.Re5 f6 >31.Re6 Qf5 32.Rxc6 Qd7 33.Ba4 > ² (0.57) Depth: 14/50 00:00:17 16285kN >24.Qxe2 Bxe2 25.Bc4 Bxc4 26.Nxc4 g4 27.f4 Qd3 28.Nxd6 cxb5 29.f5 b4 30.Bg5 bxc3 >31.bxc3 Qxc3 32.Rd1 Qc2 > -+ (-1.76) Depth: 15/63 00:03:19 195563kN >24.bxa6 Rxf2 25.Kxf2 Qh6 26.Kg1 g4 27.f4 Qf8 28.a7 Qa8 29.Nc4 Bxc4 30.Bxc4 Kg7 >31.Be3 Nd5 32.Bxd5 cxd5 33.Kg2 > ³ (-0.40) Depth: 15/63 00:05:51 332749kN > >(, 30.08.2005) > >Uri Of course, the author is not a programmer. He's just a chess player who may play w/o any computer assistance. What he says of programmes may be true for older versions. Naw machines are much better except for in closed games. Only Ruffian is the king of closed positions! Listen, how on earth you insert diagrams in posts? I wish I could do, but I still don't know how to do that.
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.