Author: Paulo Soares
Date: 15:24:39 12/10/99
Go up one level in this thread
On December 10, 1999 at 14:40:09, Dann Corbit wrote:
>On December 10, 1999 at 03:00:04, Paulo Soares wrote:
>
>>On December 09, 1999 at 20:14:01, Dann Corbit wrote:
>>
>>>On December 09, 1999 at 18:45:42, Paulo Soares wrote:
>>>>I think that are some lines that programs can't see even at
>>>>depth 20.
>>>Absolutely. Any move where the "payoff" is 10 full moves ahead will take 20
>>>plies to see. And if the payoff is farther, the computer may not make the right
>>>choice until the move is seen. Often, however, there is a simple reason why a
>>>choice was not seen. More often than not when the computer analysis is deep
>>>enough to see the benefit drawn in the human analysis, one of the following
>>>problems exists:
>>>{calling the position with a clear benefit the final board position:}
>>>0. The final board position requires an intermediate positional move.
>>>1. The final board position requires a deep sacrifice.
>>>2. There is a zugzwang position between the final board position and the
>>>current position.
>>>3. The stated move is wrong {happens more often than you might think}.
>>
>>Look for the line below, possibly the best, or less bad, for black. After 14
>>plies I placed Rebel Century (it could be another program ) to analyze the
>>position, you can see that the evaluation is low.
>>White wins this position, but in what item(0,1,2 or3)would you place the
>>position?
>>
>>1. Bxh7+ Kxh7 2. Qh5+ Kg8 3. Rg3 Bf8 4. Bg5 Qc8 5. Rc1 (5. Bf6 Qc2!?) 5...
>>Nc6 6. Bf6 Nxd4 7. Rxc8 Raxc8
>>
>>Position after 7....Raxc8:
>>2r1rbk1/1p3pp1/p3pB2/3pP2Q/P2n4/6RP/1P3PP1/6K1 w - - 0 8
>>
>>PIII-450, HT=60Mb, Default settings:
>>
>> +---+---+---+---+---+---+---+---+ ComeDann
>>8 | | | *R| | *R| *B| *K| |
>> +---+---+---+---+---+---+---+---+ 00:00:00 00:00:00
>>7 | | *p| | | | *p| *p| |
>> +---+---+---+---+---+---+---+---+
>>6 | *p| | | | *p| B | | | Time infinite
>> +---+---+---+---+---+---+---+---+ Hash 60 Mb
>>5 | | | | *p| p | | | Q | Engine Rebel Century
>> +---+---+---+---+---+---+---+---+ Personality Rebel Century
>>4 | p | | | *N| | | | | anti-GM Off
>> +---+---+---+---+---+---+---+---+ System Tiger Off
>>3 | | | | | | | R | p | Combination Off
>> +---+---+---+---+---+---+---+---+ Style Normal
>>2 | | p | | | | p | p | | Strength Strongest
>> +---+---+---+---+---+---+---+---+ Search Normal Search
>>1 | | | | | | | K | |
>> +---+---+---+---+---+---+---+---+
>> a b c d e f g h
>>
>>2r1rbk1/1p3pp1/p3pB2/3pP2Q/P2n4/6RP/1P3PP1/6K1 w - -
>>
>>00:00:00 1.00 -0.43 1.Bxg7 Bxg7
>>00:00:00 1.06 1.67 1.Rd3 gxf6 2.Rxd4
>>00:00:00 1.08 1.86 1.a5
>>00:00:00 1.11 2.08 1.b4
>>00:00:00 2.00 1.68 1.b4 Rc2 2.Rg4 Rc4 3.Rg3 Rxb4
>>00:00:00 2.05 1.71 1.b3 Rc1+ 2.Kh2
>>00:00:00 3.00 1.16 1.b3 Rc1+ 2.Kh2 Nf5 3.Rg4
>>00:00:00 3.01 1.19 1.b4 Rc1+ 2.Kh2 Rec8 3.f4
>>00:00:00 3.02 1.41 1.f4 Rc2 2.Kh2 Rxb2
>>00:00:00 3.04 1.42 1.Kh2 Nf5 2.Rg4 Rc2 3.Kg1 Rxb2
>>00:00:00 3.09 1.64 1.Rg4 Rc4 2.Rg3 Rxa4
>>00:00:00 4.00 1.54 1.Rg4 Nb3 2.g3 Rc2 3.f3 Rxb2
>>00:00:01 5.00 1.42 1.Rg4 Nf5 2.Kh2 Rc2 3.Kg1 Rxb2
>>00:00:02 6.00 1.16 1.Rg4 Nf5 2.b3 Rc1+ 3.Kh2 Rec8 4.b4
>>00:00:05 7.00 1.09 1.Rg4 Nf5 2.b4 Be7 3.b5 Bxf6 4.exf6
>> g6 5.bxa6 (3)
>>
>>00:00:08 7.12 1.14 1.Rc3 Nf5 2.Rxc8 Rxc8 3.Bg5 g6
>>00:00:10 8.00 1.09 1.Rc3 Nf5 2.Bg5 Bb4 3.Rxc8 Rxc8
>> 4.Qd1 Rc4 5.Qd3 (9)
>>
>>00:00:22 9.00 1.05 1.Rc3 Nf5 2.Bg5 Bb4 3.Rxc8 Rxc8
>> 4.Qd1 a5 5.Qb3 Nd4 6.Qd3 (18)
>>
>>00:00:57 10.00 1.11 1.Rc3 Nf5 2.Bg5 Bb4 3.Rxc8 Rxc8
>> 4.Qd1 Rc4 5.Kh2 Be7 (33)
>>
>>00:02:07 11.00 1.08 1.Rc3 Nf5 2.Bg5 Bb4 3.Rxc8 Rxc8
>>00:06:34 12.00 1.11 1.Rc3 Nf5 2.Bg5 Bb4 3.Rxc8 Rxc8
>> 4.Qd1 (196)
>>
>>
>>Positions : 99.105.320
>>Nodes per second : 131.091
>>
>>Extensions (checks) : 4.276.716 (4%)
>>Extensions (captures) : 952.553 (0%)
>>Extensions (king safety) : 354.070 (0%)
>>Extensions (on depth) : 3.858.925 (3%)
>>Extensions (remaining) : 403.564 (0%)
>>Extensions (total) : 9.845.828 (9%)
>>
>>Hash Table : 1.542.248 1.193.742 (99%)
>>Average depth : 13
>>Deepest depth : 41
>It's an interesting position.
>Crafty likes Rxg7+ with an advantage of about one pawn. Here is Crafty's
>analysis:
>2r1rbk1/1p3pp1/p3pB2/3pP2Q/P2n4/6RP/1P3PP1/6K1 w - - acd 15; acn 116978391; acs
>802; ce 94; pv Rxg7+ Bxg7 Bxg7 Kxg7 Qg4+ Kf8 Qxd4 Rc4 Qb6 Re7 Qd8+ Re8 Qd6+ Re7
>b3 Rc3 Qd8+;
>
>However, I don't think that I understand your question.
>
>Rebel settles on Rc3 with an advantage of about one pawn in about 8 seconds and
>sticks with it. It seems that both programs see a move that is hardly
>dominating (and with approximately the same advantage) and stay with it from
>there forward.
>
>Do you feel that there is some dominating move that the programs seem to miss?
>If so, what is it? I think I would have traded the bishop and the rook for the
>bishop, the pawn and the knight because the sequence is fairly obvious and it
>leaves black's king standing naked. That's just because Rebel's sequence was
>way to clever for me, besides which, I get nervous when lots of my pieces are
>under fire at the same time.
>
>I have verified that there are no sure checkmates within 9 full moves.
You didn't understand my point due to my terrible English. The position that I
placed perhaps are the best for black, starting of the initial position of Côme.
My point is that after 14 plies
(position: 2r1rbk1/1p3pp1/p3pB2/3pP2Q/P2n4/6RP/1P3PP1/6K1 w - -) ,
programs still make a low evaluation for the inicial move Bxh7, what takes me to
believe that Bxh7, although were a good move, it is not played by the programs
in view of the depth of the position. Therefore, I wrote that
programs can't see a good evaluation for Bxh7 even at depth 20.
I hope you can understand!
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.