Author: Robert Hyatt
Date: 18:23:13 07/11/00
Go up one level in this thread
On July 11, 2000 at 20:34:15, Amir Ban wrote:
>On July 11, 2000 at 18:19:36, Graham Laight wrote:
>
>>On July 11, 2000 at 15:12:00, Robert Hyatt wrote:
>>
>>>On July 11, 2000 at 11:08:27, Mark Young wrote:
>>>
>>>>On July 11, 2000 at 08:53:30, Robert Hyatt wrote:
>>>>
>>>>>On July 11, 2000 at 08:20:36, blass uri wrote:
>>>>>
>>>>>>[D]2r2rk1/4ppb1/1p2n3/4P1B1/1p6/5P2/PPP1BP2/R3K2R w KQ - 0 1
>>>>>>
>>>>>>Crafty found 22.Be3 in 24 seconds(PIII450)
>>>>>>How much time do other programs need to see Be3?
>>>>>>
>>>>>>Uri
>>>>
>>>>Crafty seems to have a very good understanding of this position, when most of
>>>>the other programs do not. Is this because you have seen this type of opposite
>>>>color bishop ending ("trap") before playing so many humans online and have taken
>>>>steps in your program to counter it.
>>>>
>>>>>
>>>>>
>>>>>Or maybe even faster: :)
>>>>>
>>>>> (3) 10 2.15 1.74 1. Bxe7 Rfe8 2. Bxb4 Nd4 3. Bd1 Nxc2+
>>>>> 4. Bxc2 Rxc2 5. f4 f6 6. Kd1 Rxb2
>>>>> 10 6.28 2.01 1. Be3 Rxc2 2. Bd3 Rxb2 3. Bh7+ Kh8
>>>>> 4. f4 b3 5. Bc1 Rxa2 6. Bb1+ Kg8 7.
>>>>> Bxa2 bxa2 8. Rxa2
>>>>> 10-> 6.78 2.01 1. Be3 Rxc2 2. Bd3 Rxb2 3. Bh7+ Kh8
>>>>> 4. f4 b3 5. Bc1 Rxa2 6. Bb1+ Kg8 7.
>>>>> Bxa2 bxa2 8. Rxa2
>>>>> 11 8.63 2.01 1. Be3 Rxc2 2. Bd3 Rxb2 3. Bh7+ Kh8
>>>>> 4. f4 b3 5. Bc1 Rxa2 6. Bb1+ Kg8 7.
>>>>> Bxa2 bxa2 8. Rxa2
>>>>> (2) 11-> 11.63 2.01 1. Be3 Rxc2 2. Bd3 Rxb2 3. Bh7+ Kh8
>>>>> 4. f4 b3 5. Bc1 Rxa2 6. Bb1+ Kg8 7.
>>>>> Bxa2 bxa2 8. Rxa2
>>>
>>>
>>>Hard to say. I definitely have code for opposite bishop positions, and I have
>>>seen crafty play such positions with some degree of 'sense'. But it is hard to
>>>say whether that is "the reason" here... I will look tonight to get an idea of
>>>what it is seeing/avoiding...
>>>
>>>In any case, humans exploit such holes all the time if you don't "fill them".
>>
>>If the loss of 0.5 points against Khalifman was avoidable by trying to manouvre
>>away from an opposite colour bishop ending, then I think that the Junior
>>programmers, while still proud to have gained such an advantage over the FIDE
>>world champion, will be kicking themselves for ultimately letting him off the
>>hook so lightly!
>>
>>-g
>
>We are not kicking ourseslves.
>
>Anyway the argument here is simplistic IMO. There are other tries for a win
>besides Be3, such as Bxb5 rather than Bd7, or Ba6 rather than Bd3. Not taking on
>g5 to avoid the queen exchange is also possible. It's just that Be3 looks like
>the best try for a win.
>
>Opposite colour bishop endings with the rooks on are quite winnable. The point
>of this one was that black's doubled pawns facing the pawn majority are not a
>gross weakness, while white's doubled f-pawns are nearly worthless.
>
>Amir
I don't think you guys ought to be kicking yourselves either. I think the
result so far is _fantastic_.
As far as opposite bishops and rooks being winnable, the chances are very
much better for a draw. This was a lesson I learned the hard way and it
took several GM conversations to convince me that what I was doing was not
weighted right. The problem is that if pawns are traded, krb vs kr is a dead
draw.. and with opposite B's, this is not easy to prevent. I have seen this
particular 'defense' used way too many times on ICC.
Another point: the "premature resignation" the other day. I have noticed
this happening _regularly_ on ICC. Players like Shirov will drop a pawn to
Crafty (or another program) and resign quickly. Where 5 years ago they would
play on and win. I think they might be a bit too early at times, but they are
seeing the methodical 'grinding' the machines can do, and tend to choose the
easy way out...
Good work, so far...
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.