Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Test Position = GM Tony Miles vs Deep Thought, 1989 [14...Bxh2!!]

Author: Robert Hyatt

Date: 12:57:24 09/09/04

Go up one level in this thread


On September 09, 2004 at 15:55:36, Michael Henderson wrote:

>On September 09, 2004 at 15:47:39, Robert Hyatt wrote:
>
>>On September 09, 2004 at 15:19:57, José Antônio Fabiano Mendes wrote:
>>
>>> GM Tony Miles vs Deep Thought, 1989
>>> [D]r4rk1/ppB2pp1/2nb3p/8/2B5/2N2P2/PP1n1P1P/2KR3R b
>>> Source ==> http://www.chesscafe.com/dvoretsky/dvoretsky.htm
>>>
>>>Now Black, it appears, must choose between 14...Nxc4 15.Bxd6 Nxd6 16.Rxd6, or
>>>14...Bxc7 15.Kxd2.In either case, he stands better; but his opponent, who has
>>>been able to maintain the material balance, has hopes of saving himself.
>>>
>>>On the other hand, several pieces are still en prise, so they’re “desperadoes,”
>>>capable of anything. Black found a completely unexpected blow, guaranteeing him
>>>a healthy extra pawn.
>>>
>>>14...Bd6xh2!! 15.Bc7xh2
>>>
>>>15.Rxd2 Bxc7 could hardly be called better. And one more “desperado” – 15.Bxf7+
>>>Rxf7 16.Bxh2 – doesn’t help: after 16...Nxf3, White is a pawn down again.
>>>
>>>15...Nd2xc4 16.Rd1-d7 b7-b6
>>>
>>>And Black eventually converted the extra pawn.
>>>
>>>Oh – and I nearly forgot to tell you who was Black. It was the computer, Deep
>>>Thought (the predecessor to Deep Blue, which beat Garry Kasparov in a match).
>>>You can understand that, for a computer, with no human habits or prejudices to
>>>shake off, finding a move like 14...Bxh2!! would not be difficult. I hope, after
>>>reading this piece, that making similar imaginative decisions will become easier
>>>for you, as well.
>>
>>The problem here is that this is a completely trivial move for a computer to
>>find.  It might have been surprising back then, considering the very slow PC
>>hardware that was around, but Crafty finds this instantly today, marking it as
>>"easy" for most any program...
>>
>
>my program Pluto finds Bxh2 at depth=1 withs score 100+, for some reason--is it
>that easy to find or do I have a bug?:
>
>--Pluto(C) v 1.10--

This is partially about evaluation, partially about tactics.  There are a couple
of ways to win a pawn.  One appears better (to me) than the other.  But early
on, Crafty flip/flops between two different captures until it gets deep enough
to realize that Bxh2 is better by some amount...



>DOS or Winboard mode?
>xboard
>setboard r4rk1/ppB2pp1/2nb3p/8/2B5/2N2P2/PP1n1P1P/2KR3R b
>analyze
>1 118 0 4055 d6h2
>2 138 0 90 d6h2 c4f7
>3 138 1 1354 d6h2 c4f7 g8f7 h1h2 d2f3
>4 128 1 4161 d6h2 d1d2 h2c7 d2d7
>5 135 4 12624 d6h2 c7h2 d2c4 d1d7 b7b6
>6 121 12 29556 d6h2 c7h2 d2c4 d1d7 b7b6 h2f4
>7 133 29 76502 d6h2 c7h2 d2c4 d1d7 b7b6 h2f4 a7a6
>8 131 68 177090 d6h2 c7h2 d2c4 d1d7 b7b6 h2f4 a8d8 d7c7
>9 132 159 391981 d6h2 h1h2 d2c4 d1d7 f8e8 h2h4 c6a5 c7a5 c4a5
>10 136 404 1095268 d6h2 h1h2 d2c4 d1d7 b7b6 f3f4 f8e8 h2h5 e8e7 h5d5 e7d7 d5d7
>11 137 948 2360233 d6h2 h1h2 d2c4 d1d7 b7b6 h2h5 f8e8 h5h4 c4e5 c7e5 c6e5
>12 133 2100 5062550 d6h2 h1h2 d2c4 d1d7 b7b6 f3f4 f8e8 b2b3 c4a5 h2h3 e8e7 h3d3
>e7d7 d3d7



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.