Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: problems of understanding connecting freepawns by computers

Author: Uri Blass

Date: 03:34:54 03/14/03

Go up one level in this thread


On March 14, 2003 at 05:54:45, Thomas Lagershausen wrote:

>On March 13, 2003 at 19:31:04, Dieter Buerssner wrote:
>
>>On March 13, 2003 at 17:16:05, Thomas Lagershausen wrote:
>>
>>>2b1r3/5pkp/6p1/4P3/QppqPP2/5RPP/6BK/8 b - - 39...c3!
>>>
>>>and the black pawns are unstoppable after 40.Qxe8? Qd7! 41.Qxd7 Bxd7
>>>
>>>I don´t know any program that finds 39...c3!
>>
>>With some patience, Yace finds c3 (after 5 1/2 min). Not with a positive score
>>yet, but I believe with more patience, the score will rise.
>>
>>It might even be not that difficult to find it much faster for my engine, by
>>givin some things a higher value. But probably, I would make the engine weaker
>>in general, while perhaps it would show some nice results in some test
>>positions.
>>
>>P4, 2.53 GHz:
>>
>>  72260313  1:33.8  -0.24 13.  1...Rd8 2. Qxb4 Qxe4 3. Qb6 Rd2 4. Rf2 Rxf2 5.
>>                               Qxf2 Qf5 6. Qd4 Be6 7. h4 {HT} {0}
>> 129437756  2:53.0  -0.17 14t  1...Rd8 2. Qxb4 Qxe4 3. Qb6 Rd3 4. Rxd3 Qxd3 5.
>>                               Qf6+ Kg8 6. Qc6 Be6 7. Qd6 Qxd6 8. exd6 c3 9.
>>                               Be4 Kf8 {0}
>> 245095369  5:39.8  -0.16 14t+ 1...c3 2. Qxe8 Qd7 3. Qxd7 Bxd7 4. Rf1 c2 5. e6
>>                               Bb5 6. Re1 b3 7. exf7 Kxf7 8. Kg1 b2 9. e5 c1=Q
>>                               10. Bf1 {421}
>> 273297580  6:22.5   0.00 14t  1...c3 2. Qxe8 Qd7 3. Qxd7 Bxd7 4. e6 Bb5 5.
>>                               Bf1 c2 6. Bxb5 c1=Q 7. e7 Qc2+ 8. Kh1 Qb1+ 9.
>>                               Kg2 Qc2+ 10. Kf1 Qd1+ 11. Kf2 Qd2+ 12. Kg1 Qc1+
>>                               13. Kg2 Qc2+ {71}
>
>These are really good news and i had the hope that Yace could handle this. But
>your statement that you will weaken your engine if it find 39..c3! quicker can´t
>be the last word.Chessknowledge means that the player understands the case that
>happened.If this pattern is not understood then it will never go over a
>invisible line.Somethings you have to know, otherwise you wil be ever in the
>darkness.


I agree
The simplest way to teach a program to find c3 faster is to increase the value
of passed pawns.

Of course it is a bad idea and the right idea is to increase the value of passed
pawns only in the right cases.



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.