Yace 0.20 versus some engines : test with black ...

[Brice Boissel]

Geschrieben von: / Posted by: Brice Boissel at 09 July 2000 01:45:07:
Results so far :
WinBoard 4.07
Athlon 550-600 MHz
Hash Tables : 16 Mb
Table-bases : 3-4-man
Pondering : off
Game in 30 minutes
Opening Book : 4.04 Mb

PROGRAMME ELO WHITE BLACK
1 Nimzo 2000 2625 0 0
2 Crafty 17.11 2620 1 1
3 LGoliath 2.8 2540 1 1
4 Comet B.23 2535 0,5 0
5 SOS 11.99 2530 1 1
6 AnMon 5.09 2520 0,5 0,5
7 Bionic 4.01 2495 0 1
8 Phalanx 22 2490 1
9 LaPetite 2490 1 0
10 TCBishop 0045 2485 1 0
11 Insomniac 0.59 2480 1 0
12 Amy 0.7 2475 1
13 Gromit 3.0 2465 0,5 0
14 Bringer 1.7 2425 1 0
15 ZChess 1.61 2415 1 0,5
16 Arasan 5.4 2360 0,5 1
17 EXchess 3.14 2340 0,5 1
18 InmiChess 3.05 2340 1 1
19 GnuChess 4.15 2300 1 1
20 Fortress 1.62 2210 0
21 SSEChess 2.02 2140 1
22 Cilian 4.13 2065 1 1
23 Sjeng 7.2 2045 1 1
24 Freyr 1.024 2040 0,5 1
* Estimation Elo = WinBoard Elo Rating 6 : please, look after Quisinsky site...
T = 57430 ; N = 24 ; M = 2393
Score with White = 18/24 (75 %) +15 =6 -3 ; Performance = 2585 Elo !
T = 48540 ; N = 20 ; M = 2427
Score with Black = 12/20 (60 %) +11 =2 -7 ; Performance = 2500 Elo
T = 105970 ; N = 44 ; M = 2408
SCORE = 30/44 (68 %) +26 =8 -10 ; PERFORMANCE "GAME IN 30 MN" = 2545 Elo
Next test : with Yace 0.21 and 6.05 Mb Mogens book...

[Dann Corbit]

Geschrieben von: / Posted by: Dann Corbit at 11 July 2000 10:06:06:
Als Antwort auf: / As an answer to: Yace 0.20 versus some engines : test with black ... geschrieben von: / posted by: Brice Boissel at 09 July 2000 01:45:07:
Out of curiosity, have you tried this experiment with any other engines?
I would expect any engine playing as black to do worse than playing as white.

My FTP site

[Brice Boissel]

Geschrieben von: / Posted by: Brice Boissel at 11 July 2000 20:47:26:
Als Antwort auf: / As an answer to: Re: Yace 0.20 versus some engines : test with black ... geschrieben von: / posted by: Dann Corbit at 11 July 2000 10:06:06:

Out of curiosity, have you tried this experiment with any other engines?
I would expect any engine playing as black to do worse than playing as white.

Hello Dan,
It' s very incredible and curious : the Yace 0.21 + Mogens book (6 Mb) :
WinBoard Plus 4.07
Athlon 550-600 MHz
Hash Tables : 16 Mb
Table-bases : 3-4-man
Pondering : off
Game in 30 minutes
Opening Book : 4.04 Mb

PROGRAMME ELO WHITE BLACK
1 Nimzo 2000 2625 1 0,5 *
2 Crafty 17.11 2620 1 1
3 Comet B.23 2540 0,5 0
4 LGoliath 2.8 2540 1 1
5 SOS 11.99 2535 1 1
6 AnMon 5.09 2515 0,5 0,5
7 Phalanx 22 2495 1 0
8 Bionic 4.01 2485 0 1
9 TCBishop 0045 2485 1 1 *
10 Gromit 3.0 2480 0,5 1 *
11 Insomniac 0.59 2475 0 0 *
12 Amy 0.7 2470 0,5 1 *
13 LaPetite 1.0 2455 0,5 1 *
14 Bringer 1.7 2430 0 0 *
15 ZChess 1.61 2425 0 1 *
16 Arasan 5.4 2370 0 1 *
17 InmiChess 3.05 2350 1 1 *
18 EXchess 3.14 2340 1 1 *
19 GnuChess 4.15 2300 1 1 *
20 Fortress 1.62 2220 1 1 *
21 SSEChess 2.02 2140 1 1 *
22 Cilian 4.13 2095 1 1 *
23 Freyr 1.024 2070 1 1 *
24 Sjeng 7.2 2050 1 1 *
* Estimation Elo = WinBoard Elo Rating 6 : please, look after Quisinsky site...
Score with White = 16,5/24 (69 %) +14 =5 -5 ; Performance = 2535 Elo
Score with Black = 19/24 (79 %) +18 =2 -4 ; Performance = 2630 Elo !
T = 115055 ; N = 48 ; M = 2397
SCORE = 35,5/48 (74 %) +32 =7 -9 ; PERFORMANCE "GAME IN 30 MN" = 2580 Elo
Now, Yace play stronger with Black !!!
Brice

[Mogens Larsen]

Geschrieben von: / Posted by: Mogens Larsen at 11 July 2000 22:42:21:
Als Antwort auf: / As an answer to: Yace 0.20 versus some engines : test with black ... geschrieben von: / posted by: Brice Boissel at 11 July 2000 20:47:26:

Out of curiosity, have you tried this experiment with any other engines?
I would expect any engine playing as black to do worse than playing as white.

Hello Dan,
It' s very incredible and curious : the Yace 0.21 + Mogens book (6 Mb) :
WinBoard Plus 4.07
Athlon 550-600 MHz
Hash Tables : 16 Mb
Table-bases : 3-4-man
Pondering : off
Game in 30 minutes
Opening Book : 4.04 Mb
PROGRAMME ELO WHITE BLACK
1 Nimzo 2000 2625 1 0,5 *
2 Crafty 17.11 2620 1 1
3 Comet B.23 2540 0,5 0
4 LGoliath 2.8 2540 1 1
5 SOS 11.99 2535 1 1
6 AnMon 5.09 2515 0,5 0,5
7 Phalanx 22 2495 1 0
8 Bionic 4.01 2485 0 1
9 TCBishop 0045 2485 1 1 *
10 Gromit 3.0 2480 0,5 1 *
11 Insomniac 0.59 2475 0 0 *
12 Amy 0.7 2470 0,5 1 *
13 LaPetite 1.0 2455 0,5 1 *
14 Bringer 1.7 2430 0 0 *
15 ZChess 1.61 2425 0 1 *
16 Arasan 5.4 2370 0 1 *
17 InmiChess 3.05 2350 1 1 *
18 EXchess 3.14 2340 1 1 *
19 GnuChess 4.15 2300 1 1 *
20 Fortress 1.62 2220 1 1 *
21 SSEChess 2.02 2140 1 1 *
22 Cilian 4.13 2095 1 1 *
23 Freyr 1.024 2070 1 1 *
24 Sjeng 7.2 2050 1 1 *
* Estimation Elo = WinBoard Elo Rating 6 : please, look after Quisinsky site...
Score with White = 16,5/24 (69 %) +14 =5 -5 ; Performance = 2535 Elo
Score with Black = 19/24 (79 %) +18 =2 -4 ; Performance = 2630 Elo !
T = 115055 ; N = 48 ; M = 2397
SCORE = 35,5/48 (74 %) +32 =7 -9 ; PERFORMANCE "GAME IN 30 MN" = 2580 Elo
Now, Yace play stronger with Black !!!
Brice

I assume this should be the 6Mb book?
Thank you for the data. It's nice to know that a week with filtering through several thousand games using CBLight wasn't completely wasted. I've made a book for Amy with the same data and the results look okay too. Thanks again.
Best wishes...
Mogens