Author: Jeremiah Penery
Date: 19:47:58 10/03/02
Go up one level in this thread
On October 03, 2002 at 22:07:38, Omid David wrote:
>On October 03, 2002 at 21:30:56, Jeremiah Penery wrote:
>
>>On October 03, 2002 at 20:39:02, Omid David wrote:
>>
>>>Quite impressive, just 1.02% slower (paractically negligible). Do you invoke
>>>that detection at every node in the search tree?
>>
>>Yes, but most of the time it exits very quickly (a couple of 'if' clauses). I
>>could probably make it cost even less, but I'm not very interested in doing that
>>yet.
>
>But the problem may arise in the endgames where there are a number of pawns and
>pieces left which can potentially cause a "block draw". Then in the draw
>detection heuristic, almost in every node you'll pass the first few 'if's and
>will have to step into a detailed assessment of the position. How much does it
>slow down there, in comparison to the non-detecting version?
I'm not really sure, and I suspect it depends highly on the kind of position.
If you have any specific ones, I'll be happy to test them.
I did run one of the positions I posted before (the last one), because it seemed
like one where the routine would make a lot of difference.
4k3/8/1p1p4/pPpPp1p1/P1P1PpPp/5P1P/2BB4/K7 w
WITHOUT the heuristic:
22 7.09 7.89 1. Kb2 Ke7 2. Bc1 Kd8 3. Bd1 Kd7 4.
Kc2 Ke7 5. Kd2 Kf7 6. Bb3 Kf6 7. Ba3
Kg6 8. Bb2 Kf7 9. Ke2 Ke8 10. Kd1 Ke7
11. Bc3 Kd7 <HT>
22-> 8.51 7.89 1. Kb2 Ke7 2. Bc1 Kd8 3. Bd1 Kd7 4.
Kc2 Ke7 5. Kd2 Kf7 6. Bb3 Kf6 7. Ba3
Kg6 8. Bb2 Kf7 9. Ke2 Ke8 10. Kd1 Ke7
11. Bc3 Kd7 <HT>
23 12.50 7.89 1. Kb2 Ke7 2. Bc1 Kd8 3. Bd1 Kd7 4.
Kc2 Ke7 5. Kd2 Kf7 6. Bb3 Kf6 7. Ba3
Kg6 8. Bb2 Kf7 9. Ke2 Ke8 10. Bc3 Ke7
11. Bb2 <HT>
23-> 19.00 7.89 1. Kb2 Ke7 2. Bc1 Kd8 3. Bd1 Kd7 4.
Kc2 Ke7 5. Kd2 Kf7 6. Bb3 Kf6 7. Ba3
Kg6 8. Bb2 Kf7 9. Ke2 Ke8 10. Bc3 Ke7
11. Bb2 <HT>
24 23.79 7.89 1. Kb2 Ke7 2. Bc1 Kd8 3. Bd1 Kd7 4.
Kc2 Ke7 5. Kd2 Kf7 6. Bb3 Kf6 7. Ba3
Ke7 8. Kc3 Kd7 9. Bb2 Ke7 10. Kd2 Kd7
11. Bc3 Kd8 12. Kd3 Kd7
24-> 29.88 7.89 1. Kb2 Ke7 2. Bc1 Kd8 3. Bd1 Kd7 4.
Kc2 Ke7 5. Kd2 Kf7 6. Bb3 Kf6 7. Ba3
Ke7 8. Kc3 Kd7 9. Bb2 Ke7 10. Kd2 Kd7
11. Bc3 Kd8 12. Kd3 Kd7
Elapsed Time: 29.96, CPU Time: 29.29, CPU Use: 97%
Material Balance: 700, Previous Moves Predicted: 0
Nodes: 35959245, Evaluations: 904616, FH%: 96%, NPS: 1200k
EXTENSIONS->
Threat: 3412, Passed Pawn: 678114, Recapture: 14914
One-Reply: 64253, Check: 1799411
PRUNING-> Razors-> White: 375133, Black: 0
Futility-> White: 7870027, Black: 13994
TABLEBASES-> Probes: 0, Successful: 0
HASH TABLES-> Hash Hits: 84% (15042249/17741754)
Pawn Hash Hits: 99% (13954512/13969872)
Hash Table Filled: 92%
WITH the heuristic:
22 8.30 6.56 1. Kb2 Kd7 2. Kc3 Kc7 3. Kd3 Kd7 4.
Be1 Ke7 5. Bb3 Kd7 6. Bc3 Ke7 7. Kc2
Kd7 8. Bd2 Ke7 9. Be1 Kd7 10. Kd3 Ke7
11. Bxa5 bxa5
22-> 9.64 6.56 1. Kb2 Kd7 2. Kc3 Kc7 3. Kd3 Kd7 4.
Be1 Ke7 5. Bb3 Kd7 6. Bc3 Ke7 7. Kc2
Kd7 8. Bd2 Ke7 9. Be1 Kd7 10. Kd3 Ke7
11. Bxa5 bxa5
23 12.83 6.56 1. Kb2 Kd7 2. Kc3 Kc7 3. Kd3 Kd7 4.
Be1 Ke7 5. Bb3 Kd7 6. Bc3 Ke7 7. Kc2
Kd7 8. Bd2 Ke7 9. Kd3 Kd7 10. Bc2 Ke7
11. Bxa5 bxa5 12. Bb3
23-> 14.01 6.56 1. Kb2 Kd7 2. Kc3 Kc7 3. Kd3 Kd7 4.
Be1 Ke7 5. Bb3 Kd7 6. Bc3 Ke7 7. Kc2
Kd7 8. Bd2 Ke7 9. Kd3 Kd7 10. Bc2 Ke7
11. Bxa5 bxa5 12. Bb3
24 24.65 6.55 1. Kb2 Kd7 2. Kc3 Kc7 3. Kd3 Kd7 4.
Be1 Ke7 5. Bb3 Kd7 6. Kc2 Ke7 7. Kd2
Kd7 8. Bc2 Ke8 9. Bd3 Kd7 10. Bf2 Ke7
11. Kc3 Kd8 12. Bxc5 dxc5
24-> 28.04 6.55 1. Kb2 Kd7 2. Kc3 Kc7 3. Kd3 Kd7 4.
Be1 Ke7 5. Bb3 Kd7 6. Kc2 Ke7 7. Kd2
Kd7 8. Bc2 Ke8 9. Bd3 Kd7 10. Bf2 Ke7
11. Kc3 Kd8 12. Bxc5 dxc5
Elapsed Time: 28.13, CPU Time: 27.40, CPU Use: 97%
Material Balance: 700, Previous Moves Predicted: 0
Nodes: 30677191, Evaluations: 1131204, FH%: 97%, NPS: 1090k
EXTENSIONS->
Threat: 2572, Passed Pawn: 485145, Recapture: 13150
One-Reply: 43861, Check: 1342909
PRUNING-> Razors-> White: 240096, Black: 0
Futility-> White: 5085517, Black: 15205
TABLEBASES-> Probes: 0, Successful: 0
HASH TABLES-> Hash Hits: 99% (15940630/16061384)
Pawn Hash Hits: 99% (10896693/10907130)
Hash Table Filled: 85%
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.