Author: Miguel A. Ballicora
Date: 14:48:40 05/09/02
Go up one level in this thread
On May 08, 2002 at 16:55:57, Matthias Gemuh wrote:
>On May 08, 2002 at 16:26:18, Miguel A. Ballicora wrote:
>
>>On May 08, 2002 at 13:04:58, Matthias Gemuh wrote:
>>
>>>
>>>
>>>I mean those that cause a return.
>>>
>>>Matthias.
>>
>>I think that mine is lower, more like 10-15% on the middlegame.
>>It is lower because I do store and probe in the qsearch?
>>
>>Maybe we can chose a set of positions, analyze them to a fix depth
>>an compare numbers.
>>
>>Regards,
>>Miguel
>
>
>Would many other parameters like nullmove, internal iter. deepening,
>aspiraration windows, move ordering, extensions, fut. pruning, pawn tables, etc
>not affect the meaning of the results ?
>
>Send me positions and say how to proceed.
>Send to xxxx@online.de, but replace "xxxx" with "mgemuh".
>
>Regards,
>Matthias.
These are the positions I chose to run for 9 plies each.
rnr3k1/p3qpp1/1p2b2p/1Bpp4/3P4/Q3PN2/PP3PPP/2R1K2R b K - bm Qb7;id "DCS.001";
2r2qk1/r2n2p1/p3p2p/2pp4/8/Q3P3/PP2BPPP/2R2RK1 w - - bm e4;id "DCS.002";
4q2k/2r1r1p1/4Pn1p/p1p2R2/P2pQ2P/1P1B1R2/6P1/6K1 w - - bm Rxf6;id "DCS.003";
r1b1r1k1/pp3ppp/2p4q/3pP2n/3P4/2PB1P1P/P1Q1N1P1/R3K2R w KQ - bm Qd2;id
"DCS.004";
2r2rk1/1q1bbp1p/n2p1np1/1p1Pp3/1P2P3/4NN1P/3B1PP1/RB1QR1K1 w - - bm Bd3;id
"DCS.005";
r2q1rk1/4bppp/p1b1pn2/1p1p4/3P1B2/1QN3PP/PP2PPB1/2R2RK1 b - - bm Qd7;id
"DCS.006";
2r2rk1/pp1bRppp/1q3n2/n2p4/3P1B2/3B1N2/PP3PPP/2RQ2K1 w - - bm Ng5;id "DCS.007";
2rn2k1/p2qnrpp/1p2p3/3pPp1P/1B1P1N2/P2Q2R1/2P2PP1/R3K3 b - - bm Rc4;id
"DCS.008";
2r3k1/1b2ppb1/q2p1np1/2p3N1/1p2P1PQ/nP1PB2P/r1P1N1B1/2R2RK1 w - - bm Rxf6;id
"DCS.009";
rn1qrbk1/1pp2ppp/8/p2bp3/Q7/P2PBNP1/1PR1PPBP/2R3K1 w - - bm Rxc7;id "DCS.010";
Attempts Hits (%)
Partial Perfect
Probe in quies search
1 7987553 20.2 13.8
2 9549038 14.2 10.3
3 2659876 21.6 13.3
4 3459439 19.3 12.9
5 4987247 16.0 10.9
6 9928127 17.2 12.2
7 16195160 17.5 12.4
8 2137209 18.5 11.9
9 10904976 12.7 9.0
10 6871937 17.8 12.1
--------------------------
Av. 7468056 17.5 11.9
Probe only on regular search
1 3026432 32.3 23.1
2 4040161 24.2 18.0
3 1470444 29.8 18.9
4 1979177 29.2 20.5
5 2845284 26.8 18.9
6 4793760 25.7 19.1
7 9269009 26.0 19.4
8 1309214 29.0 19.2
9 5480272 21.1 15.1
10 3365744 29.1 20.9
--------------------------
Av 3757950 27.3 19.3
These numbers are more similar to yours, probably. That indicates that
to compare the % of hits it should be taken into account if the program
probes in the quiescent search.
Regards,
Miguel
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.