Author: Richard A. Fowell (fowell@netcom.com)
Date: 21:39:00 11/26/99
Go up one level in this thread
On November 24, 1999 at 22:42:52, yibing fan wrote:
>This position was rearched after white's 15th move in a game between
>Duz_Hotimirsky F I - Yudovich Mikhail M, 1948. Black to move. The game only
>lasted 20 moves.
>
>3r1rk1/p2q1ppp/2pb1n2/5b2/1nB5/2N1BN1P/PP2QPP1/R4RK1 b - - id=Duz_Hotimirsky F I
> - Yudovich Mikhail M ; bm=f5h3;
>
>All my top chess softwares failed to find the best move BxP(h3) within 8 hrs
>search on a PII 266. I tried crafty 16.6, Rebel 8, CM6000. Try yours! It is a
>good position to keep your chess engine busy for whole Thanksgiving holidays!
>
<snip>
On my 180 MHz Mac:
HIARCS 7 picked Bxh3 in 4 min 35 seconds,
MacChess 5.01 picked Bxh3 in 5 hr 6 minutes.
Times would be a lot faster on a faster Mac (e.g., the current 400 MHz iMac).
One interesting thing I notice, though, is that for both HIARCS and MacChess,
it takes considerably longer before they evaluation Bxh3 as significantly
better than Rfe8. For example, if Rfe8 is forbidden to MacChess, it finds
Bxh3 in 20 minutes, with the same rating as Rfe8 had. Also, it takes HIARCS
more than 1 hour and 10 minutes to rate Bxh3 more than 0.1 pawn better than
Rfe8.
(both of these checked using the "next best move" features of these programs.
Note that, while this position does not seem to be as tough as a Nolot position,
the programs that solve it quickly may not see Bxh3 as the significantly best
move for a while - HIARCS didn't, anyway.
I've appended the log files.
Richard A. Fowell (fowell@netcom.com)
=========== HIARCS 7.0 Mac : 4 min 35 sec ===================
8»27 275 4217kN f5xh3 g2xh3 d7xh3 e3xa7 h3-g4+ g1-h1 f8-e8 a7-e3 = 53
9 2462 36mN f5xh3 g2xh3 d7xh3 a1-b1 f6-g4 c3-e4 g4xe3 = 90
10 4296 63mN f5xh3 g2xh3 d7xh3 a1-d1 f6-g4 d1xd6 d8xd6 e3-g5 = 88
11 1...Bxh3 gxh3 Qxh3 Rad1 Ng4 Rxd6 Rxd6 Bg5 Nd5 Ne4 Re6 = 88
on "next best move":
8»27 63 859kN b4-d5 c4xd5 f6xd5 c3xd5 c6xd5 a1-c1 f8-e8 e2-a6 = 43
9»29 451 6027kN b4-d5 c4xd5 f6xd5 c3xd5 c6xd5 f1-d1 f8-e8 e2-a6 = 44
9»29 676 8822kN f8-e8 a1-d1 b4-d5 c3xd5 f6xd5 f3-h4 f5-e6 c4xd5 = 53
10 1849 22mN f8-e8 f1-c1 b4-d5 c3xd5 f6xd5 c4xd5 c6xd5 e2-a6 = 56
11 7336 90mN f8-e8 f1-c1 b4-d5 c3xd5 f6xd5 c4xd5 c6xd5 f3-d4 = 54
12 1...Rfe8 Rfc1 Nbd5 Nxd5 Nxd5 Bxd5 cxd5 Nd4 Rb8 Qd2 = 54
on "next best move" again:
9»29 260 3573kN b4-d5 c3xd5 f6xd5 f1-e1 f8-e8 c4xd5 c6xd5 = 47
11 7942 101mN b4-d5 c3xd5 f6xd5 f1-e1 f8-e8 c4xd5 c6xd5 f3-d4 = 50
12 1...Nbd5 Nxd5 Nxd5 Rfe1 Rfe8 Bxd5 cxd5 Nd4 Be4 Rac1 = 50
=========== MacChess 5.01 Mac: 5 hr 6 min - close in 20 min. ===================
MacChess 5.01 takes a little under 6 hours this move on my
180 MHz Mac, though, looking at the score, one might argue that it takes
one more ply (just under 11 hours) to begin to understand the significance
of it's find. Here's the (slightly cleaned up) log.
I see in other posts that recent Rebel versions and HIARCS do fine on this one.
I expected CM6000 to do better than your results, but we have seen a large
scatter on CCC test positions like this.
Ply Hr:Mn:Sec Nodes Score Principal Variation (clipped, save last 2)
==============================================================================
06/01|00:00:01| 122692| +3| Nc2 Rac1 Nxe3 Qxe3 Rfe8 Qd4 a5
07/01|00:00:03| 457901| +4| Nc2 Rac1 Nxe3 Qxe3 Rfe8 Qg5 Ne4 Qe3
08/01|00:00:14| 2219942| +3| Nc2 Rad1 Rfe8 Nh4 Nxe3 fxe3 Be6 Bxe6
09/01|00:00:42| 6577295| +5| Nc2 Rad1 Rfe8 Nh4 Nxe3 fxe3 Be6 Bxe6
10/01|00:02:31| 24033303| -2| Nc2 Rad1 Rfe8 Nh4 Nxe3 fxe3 Be6 Rfe1
10/02|00:09:45| 94285133| +0| Rfe8 Nh4 Bxh3 gxh3 Qxh3 Nf3 Nxa2 Bxa2
11/01|00:43:17| 418444895| +2| Rfe8 Rac1 Nfd5 Nxd5 Nxd5 Bxd5 cxd5 Nd4
12/01|02:23:13| 1392684381| +5| Rfe8 Rac1 Bf4 Rfd1 Nbd5 Nxd5 Nxd5 Bxd5
12/43|05:05:53| 2922205053| +22| Bxh3 a3 Qg4 Ne1 Nbd5 Qxg4 Bxg4 Bxd5
Nxd5 Nxd5 cxd5 Bxa7 Rfe8
13/01|10:44:06| 6120317215| +86| Bxh3 Rfd1 Bg4 Bf4 Nbd5 Bxd6 Qxd6 Ne4
Qf4 Ned2 Rfe8 Qf1 a5
If we ask MacChess to find the "next best move" once it has selected Rfe8, it
finds Bxh3 relatively quickly (20 minutes at 10 ply), and rates it about
the same as Rfe8.
(MacChess 5.01 analysis of Position with Rfe8 excluded from consideration)
Ply Hr:Mn:Sec Nodes Score Principal Variation (clipped, save last 2)
==============================================================================
06/01|00:00:01| 122336| +3| Nc2 Rac1 Nxe3 Qxe3 Rfe8 Qd4 a5
07/01|00:00:03| 413309| +4| Nc2 Rac1 Nxe3 Qxe3 Rfe8 Qg5 Ne4 Qe3
08/01|00:00:14| 2113216| +3| Nc2 Rad1 Rfe8 Nh4 Nxe3 fxe3 Be6 Bxe6
09/01|00:00:43| 6315024| +5| Nc2 Rad1 Rfe8 Nh4 Nxe3 fxe3 Be6 Bxe6
10/01|00:02:38| 23709967| -2| Nc2 Rad1 Rfe8 Nh4 Nxe3 fxe3 Be6 Rfe1
10/42|00:19:48| 175666619| +0| Bxh3 gxh3 Qxh3 Rad1 Nxa2 Nxa2 Qg4+
Kh1 Qh3+ Kg1
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.