Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: A different kind of test.

Author: Slater Wold

Date: 21:34:53 08/08/02

Go up one level in this thread


On August 09, 2002 at 00:17:33, John Merlino wrote:

>On August 08, 2002 at 23:20:47, Slater Wold wrote:
>
>>[D]3r1brk/1p3p1p/p1npp1p1/3Nn1Q1/P1P1PN2/1q4R1/5PPP/B1R3K1 b - - 0 24
>>
>>How long does it take your engine to understand it's losing?
>>
>>How long does it take it to understand it's losing badly?
>
>It takes Chessmaster 9000, on a PIII-733, 4:14 to see that Qxa4 loses. The fail
>low doesn't resolve until 7:39, in favor of Qa2, which doesn't look TOO bad at
>first sight. On second sight, it looks a little worse:
>
>Time	Depth	Score	Positions	Moves
>0:00	1/3	-1.32	3722		24...Qxa4 25.Nb6 Qb4 26.Bd4
>0:00	1/4	-1.67	13895		24...Qxa4 25.Nf6 Be7 26.Qh4 Rg7
>0:00	1/5	-1.44	37160		24...Qxa4 25.Nf6 Be7 26.Qh4 Rg7
>					27.Rh3
>0:01	1/6	-1.36	109217		24...Qxa4 25.Nf6 Be7 26.Qh4 Rg7
>					27.Rh3 Bxf6 28.Qxf6
>0:04	1/7	-1.41	322890		24...Qxa4 25.Nf6 Be7 26.Qh4 Rg7
>					27.Rh3 Bxf6 28.Qxf6 Re8 29.Bb2 Qa5
>0:10	1/8	-1.38	859511		24...Qxa4 25.Nf6 Be7 26.Qh4 Rg7
>					27.Rh3 Bxf6 28.Qxf6 Qa2 29.Nd3
>					Qd2 30.Rb1
>0:27	1/9	-1.29	2430899		24...Qxa4 25.Nf6 Be7 26.Qh4 Rg7
>					27.Rh3 Bxf6 28.Qxf6 Re8 29.Rd3
>					Qb4 30.Rcd1 Qxc4 31.Rxd6
>1:13	1/10	-1.09	6848593		24...Qxa4 25.Nf6 Be7 26.Qh4 Rg7
>					27.Rh3 Bxf6 28.Qxf6 Qa5 29.Rd1
>					Kg8 30.Ne2 Re8 31.Bxe5 Nxe5
>4:14	2/11	1.48	23163907	24...Qxa4 25.Nf6 Be7 26.N4h5 Rg7
>					27.Qh6 Rdg8 28.f4 Ng4 29.Rxg4 Qa3
>					30.Rd1 Qb3
>7:39	2/11	0.60	42138807	24...Qa2 25.Nf6 Be7 26.N4h5 Rg7
>					27.Qh6 Rdg8 28.Bc3 Bd8 29.Nxg7
>					Bxf6 30.Nh5
>15:32	3/12	1.04	88344527	24...Qa2 25.Nf6 Be7 26.N4h5 Rg7
>					27.Qh6 Rdg8 28.Ne8 Bh4 29.Nhxg7
>					Bxg3 30.hxg3 Qxa4 31.Nh5
>
>jm

Considering the HW, that's not too bad.  With a 1.5Ghz you could probably bring
it down to 2 minutes or so.

Of course, this game is all but over, so 1.04 isn't exactly seeing it's lost.
You can still draw (and sometimes win) with a 1.04 score against other
computers.  However, this is from a computer vs 2550 GM and it's *over*.

If you play Qa2 Nf6 and let it think, Be7 will fail low as well, probably past
4.00+ or so.  Then it will decide to lose the rook, and it will even out the
rest of the game.



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.