wacnew.epd & single search improvements

[Stefan Knappe]

Geschrieben von: / Posted by: Stefan Knappe at 16 June 2003 18:00:47:

Hi,
last week I started to research the effect of isolated improvements of my program (Matador) on the playing strength of the program. Perhaps someone of you is also interested in the results?! However I am interested in your opinion about the test and the results!
At first I compiled a version of Matador without any improvements of the search. This version worked without any kind of sorting, extensions, pruning, hash … The evalution based exclusively on material.
Okay, at first the results of Matador on a 466 Mhz Celeron with 33 MB hash and all activated technics:
//Test: wacnew.epd / Hash: 33.0MB / Time: 5s
total solve time : 251.28 s / solved: 269/300
avrg. depth: 8.6 / avrg. max depth: 22.9
avrg. QS: 10.8% / avrg. MO: 92.0%
The results of the following test (pure alpha-beta algo with zero window and quiescence search) look a little bit poore. But I assume, it is not realistic to expect better results?! Did you expect better results?
//Test: wacnew.epd / Hash: 33.0MB / Time: 5s
total solve time : 952.17 s / solved: 115/300
avrg. depth: 5.5 / avrg. max depth: 21.6
avrg. QS: 77.8% / avrg. MO: 61.7%
In my opinion the most interesting figures are the average QS-nodes of 77,8% (compared to the 10,8% of the test before) and the average move ordering of 61,7%. I expected a more bad ordering. Probably there are a lot of senseless moves, which will be disproved by each next move?!
Now I will test the effect of the nullmove (R=2). What effect do you expect?
Regards,
Stefan

[Andreas Herrmann]

Geschrieben von: / Posted by: Andreas Herrmann at 17 June 2003 10:56:05:
Als Antwort auf: / In reply to: wacnew.epd & single search improvements geschrieben von: / posted by: Stefan Knappe at 16 June 2003 18:00:47:

Hi,
last week I started to research the effect of isolated improvements of my program (Matador) on the playing strength of the program. Perhaps someone of you is also interested in the results?! However I am interested in your opinion about the test and the results!
At first I compiled a version of Matador without any improvements of the search. This version worked without any kind of sorting, extensions, pruning, hash … The evalution based exclusively on material.
Okay, at first the results of Matador on a 466 Mhz Celeron with 33 MB hash and all activated technics:
//Test: wacnew.epd / Hash: 33.0MB / Time: 5s
total solve time : 251.28 s / solved: 269/300
avrg. depth: 8.6 / avrg. max depth: 22.9
avrg. QS: 10.8% / avrg. MO: 92.0%
The results of the following test (pure alpha-beta algo with zero window and quiescence search) look a little bit poore. But I assume, it is not realistic to expect better results?! Did you expect better results?
//Test: wacnew.epd / Hash: 33.0MB / Time: 5s
total solve time : 952.17 s / solved: 115/300
avrg. depth: 5.5 / avrg. max depth: 21.6
avrg. QS: 77.8% / avrg. MO: 61.7%
In my opinion the most interesting figures are the average QS-nodes of 77,8% (compared to the 10,8% of the test before) and the average move ordering of 61,7%. I expected a more bad ordering. Probably there are a lot of senseless moves, which will be disproved by each next move?!
Now I will test the effect of the nullmove (R=2). What effect do you expect?
Regards,
Stefan

Hi Stefan,
interessting test. I will do that also in the next weeks. But i want to enable/disable all parts by different compiler defines. So it's a little bit work. A few things i can at the moment enable/disable by compiler defines, but not all.
I'm anxious about your results with nullmove R=2.
best wishes
Andreas