Author: Vincent Diepeveen
Date: 05:42:50 03/03/01
Go up one level in this thread
On March 02, 2001 at 08:51:00, Robert Hyatt wrote:
>On March 02, 2001 at 07:21:29, Vincent Diepeveen wrote:
>
>>On March 02, 2001 at 02:14:20, Uri Blass wrote:
>>
>>>On March 02, 2001 at 00:20:20, Albert Silver wrote:
>>>
>>>>On March 01, 2001 at 22:31:24, Albert Silver wrote:
>>>>
>>>>>You should look below. Uri has shed some doubt on the draw.
>>>>>
>>>>> Albert
>>>>
>>>>It's a draw, nevermind.
>>>
>>>It was not a draw at least in the game between chessmaster8000 and itself
>>>see http://www.icdchess.com/forums/1/message.shtml?156697
>>>
>>>I did not see a forced line that lead to a draw and the position should be
>>>analyzed to prove if it is a draw or not a draw.
>>>
>>>Uri
>>
>>DIEP says it's a draw and i go for DIEP instead of a forward pruning
>>prog called The King, which probably played on a level similar to 5 0.
>>Did it play Qe3 at the first move anyway?
>>
>>Also Seirawan shows in june 1997 ICCA journal that it's a draw, besides
>>that i did some analysis myself in 1997 and then also concluded it
>>was a draw.
>>
>>So my friend, where your analysis usually are there before anyone
>>has said a word, here you argue lotta GMs and an objective chess prog
>>without anything, how comes?
>
>
>Sorry, but I don't believe you can find a forced draw here. There are way too
>many very deep but quiet moves that can be played. Including the option by
>white of simply giving up the bishop to get the passed pawn moving. A program
>might think that white is losing there. It takes one deep search to figure this
>out.
>
>Crafty gets a draw score at depth=16 for Qe3. But it then loses it at depth=17
>when it realizes that one side can do better. 0.00 doesn't impress me at all
>here without the full 60+ ply variation for the deepest forced draw.
Deep Blue didn't even expect Qe3!
What do i need to say more?
In diep the low score doesn't go away. Crucial to see are a few
checking lines. In DIEP i can walk through hash and i see that
the entire line as quoted by Seirawan to be drawn, including h4 h5 move,
is seen by DIEP, with a score 0.00.
In diep there is only 1 score 0.00 and that's repetition, 50 move rule or
stalemate...
Also Deep Blue has a very pessimistic score for *any* exchanging the
queen line. All these lines are bigtime won for white.
Where i get a score of like 3 pawns, deep blue seems to give a huge
penalty for opposite bishops *whatever* the position. But still it's
1.25 pawn or something (assuming a pawn in deep blue is 128, not
sure whether that's true).
Seeing the tactical draw as written down by Seirawan at page 110 under
a:
45..Qe3 46. Qxd6,Re8 47.h4,h5 etcetera
This 46..Re8 line is crucial for the machine to see as a draw, as
that means the whole score of the line is getting near zero then, depending
upon how you evaluate opposite bishops in the side lines.
We know clearly that Deep Blue penalized opposite bishops bigtime, so
obviously it only needed to see the Qe3 + Re8 lines to give it a near
to draw score.
Obviously it didn't see it!
Vincent
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.