Page 3 of 8
Re: A dangerous combination
Posted: Fri Mar 30, 2007 9:15 pm
by Tony Thomas
Code: Select all
4 RomiChessDK5 : 2489 181 (+ 94,= 24,- 63), 58.6 %
Danasah 2.85 : 30 (+ 16,= 6,- 8), 63.3 %
Phalanx Reborn : 30 (+ 15,= 3,- 12), 55.0 %
Francesca MAD 0.13 : 30 (+ 10,= 4,- 16), 40.0 %
Lime 6.3 : 30 (+ 27,= 2,- 1), 93.3 %
Zappa 1.1 : 30 (+ 10,= 4,- 16), 40.0 %
GreKo 5.2 : 1 (+ 0,= 1,- 0), 50.0 %
Arasan 9.5 : 30 (+ 16,= 4,- 10), 60.0 %
Performance against Phalanx not as good as expected, but still not too bad. I hope she will continue to perform as good as it did so far.
Re: A dangerous combination
Posted: Fri Mar 30, 2007 9:25 pm
by Michael Sherwin
Tony Thomas wrote:Code: Select all
4 RomiChessDK5 : 2489 181 (+ 94,= 24,- 63), 58.6 %
Danasah 2.85 : 30 (+ 16,= 6,- 8), 63.3 %
Phalanx Reborn : 30 (+ 15,= 3,- 12), 55.0 %
Francesca MAD 0.13 : 30 (+ 10,= 4,- 16), 40.0 %
Lime 6.3 : 30 (+ 27,= 2,- 1), 93.3 %
Zappa 1.1 : 30 (+ 10,= 4,- 16), 40.0 %
GreKo 5.2 : 1 (+ 0,= 1,- 0), 50.0 %
Arasan 9.5 : 30 (+ 16,= 4,- 10), 60.0 %
Performance against Phalanx not as good as expected, but still not too bad. I hope she will continue to perform as good as it did so far.
63.3 % vs DanaSah 2.85 is huge for Romi !
Re: A dangerous combination
Posted: Fri Mar 30, 2007 9:33 pm
by Tony Thomas
Michael Sherwin wrote:Tony Thomas wrote:Code: Select all
4 RomiChessDK5 : 2489 181 (+ 94,= 24,- 63), 58.6 %
Danasah 2.85 : 30 (+ 16,= 6,- 8), 63.3 %
Phalanx Reborn : 30 (+ 15,= 3,- 12), 55.0 %
Francesca MAD 0.13 : 30 (+ 10,= 4,- 16), 40.0 %
Lime 6.3 : 30 (+ 27,= 2,- 1), 93.3 %
Zappa 1.1 : 30 (+ 10,= 4,- 16), 40.0 %
GreKo 5.2 : 1 (+ 0,= 1,- 0), 50.0 %
Arasan 9.5 : 30 (+ 16,= 4,- 10), 60.0 %
Performance against Phalanx not as good as expected, but still not too bad. I hope she will continue to perform as good as it did so far.
63.3 % vs DanaSah 2.85 is huge for Romi !
I hope that we can get the engine to the 3rd division by next edition if not to second. I know that Leo will promote an engine directly two divisions up if it scores well enough to get a rating that high. In the case of Romi, I doubt that she can pull off such a high score.
Re: A dangerous combination
Posted: Fri Mar 30, 2007 9:50 pm
by Dr.Wael Deeb
Tony Thomas wrote:Michael Sherwin wrote:Tony Thomas wrote:Code: Select all
4 RomiChessDK5 : 2489 181 (+ 94,= 24,- 63), 58.6 %
Danasah 2.85 : 30 (+ 16,= 6,- 8), 63.3 %
Phalanx Reborn : 30 (+ 15,= 3,- 12), 55.0 %
Francesca MAD 0.13 : 30 (+ 10,= 4,- 16), 40.0 %
Lime 6.3 : 30 (+ 27,= 2,- 1), 93.3 %
Zappa 1.1 : 30 (+ 10,= 4,- 16), 40.0 %
GreKo 5.2 : 1 (+ 0,= 1,- 0), 50.0 %
Arasan 9.5 : 30 (+ 16,= 4,- 10), 60.0 %
Performance against Phalanx not as good as expected, but still not too bad. I hope she will continue to perform as good as it did so far.
63.3 % vs DanaSah 2.85 is huge for Romi !
I hope that we can get the engine to the 3rd division by next edition if not to second. I know that Leo will promote an engine directly two divisions up if it scores well enough to get a rating that high. In the case of Romi, I doubt that she can pull off such a high score.
One division up the rating list will be a good achievement,I don't want the bitch to get overheated
Why do you talked about this stuff in public ?
Posted: Fri Mar 30, 2007 11:11 pm
by Daniel Mehrmann
This whole thread is just about internal testing of romi books.
That's great of course, i'm a fan of romi, but why do you do this stuff in public ???
I'm not interested and i guess a lot of people as well.
Maybe posting the final results is enought. All other stuff should be done between developer and tester via e-mail as all other developers and tester guys doing it too !! And it's not Michael fault here....
Is it soooooo difficult ?
Best,
Daniel
Re: Why do you talked about this stuff in public ?
Posted: Sat Mar 31, 2007 12:29 am
by Michael Sherwin
Daniel Mehrmann wrote:This whole thread is just about internal testing of romi books.
That's great of course, i'm a fan of romi, but why do you do this stuff in public ???
I'm not interested and i guess a lot of people as well.
Maybe posting the final results is enought. All other stuff should be done between developer and tester via e-mail as all other developers and tester guys doing it too !! And it's not Michael fault here....
Is it soooooo difficult ?
Best,
Daniel
Tony has tested the last couple of Romi betas under the same conditions (including the same book). His excitement (and mine) is that this beta seems to have taken another big jump in strength. The opening book talk is just side talk and not the point of the thread. But, I understand that the side talk is not interesting to most!
Re: Why do you talked about this stuff in public ?
Posted: Sat Mar 31, 2007 7:52 am
by Dr.Wael Deeb
Daniel Mehrmann wrote:This whole thread is just about internal testing of romi books.
That's great of course, i'm a fan of romi, but why do you do this stuff in public ???
I'm not interested and i guess a lot of people as well.
Maybe posting the final results is enought. All other stuff should be done between developer and tester via e-mail as all other developers and tester guys doing it too !! And it's not Michael fault here....
Is it soooooo difficult ?
Best,
Daniel
And what is your problem particulary,what did we do wrong to disturbe your inner state of peace?
If you're not interested,skip the post,simple as that
Re: Why do you talked about this stuff in public ?
Posted: Sun Apr 01, 2007 5:44 am
by Tony Thomas
Daniel Mehrmann wrote:This whole thread is just about internal testing of romi books.
That's great of course, i'm a fan of romi, but why do you do this stuff in public ???
I'm not interested and i guess a lot of people as well.
Maybe posting the final results is enought. All other stuff should be done between developer and tester via e-mail as all other developers and tester guys doing it too !! And it's not Michael fault here....
Is it soooooo difficult ?
Best,
Daniel
I have only tested one book and that's all. Michael adviced me that if I used a book made from grand07.pgn in combination with the current book, it will help his engine score an additional 8% or so. I responded
Also, you know that changing the books a lot can skew the rating a lot, you wnt be able to tell weather it was the engine or the book
May be you wish you had a tester like me. After all this is the tournament forum and here I can post about any kind of test involving a chess engine. People posts results of 100 different versions of Rybka, using 100 different books, and 100 different setttings, you dont seem to be questioning them.
[Edited by moderation]
Re: A dangerous combination
Posted: Mon Apr 02, 2007 8:39 am
by Tony Thomas
The rating of Romi has dropped a lot, but here are the current standings....Yes, she is still at 4th place. I guess I am gonna have to pick new opponents as she out performs about 75% of her 12 opponents. I am gonna have to let go of most of the opponents towards the end of the list and find new ones that are stronger than Zappa. I might even add a commercial engine.
Code: Select all
4/2/2007 12:56:26 AM :
Program Elo + - Games Score Av.Op. Draws
1 Zappa 1.1 : 2543 29 29 540 70.0 % 2396 16.7 %
2 Francesca MAD 0.13 : 2501 28 27 540 64.4 % 2398 18.3 %
3 Arasan 9.5 : 2496 28 28 540 63.6 % 2399 16.5 %
4 RomiChessDK5 : 2466 35 35 330 59.8 % 2397 16.7 %
5 Danasah 2.85 : 2444 26 26 541 56.1 % 2402 22.7 %
6 RomiChessDK3 : 2444 33 33 360 57.4 % 2392 18.1 %
7 RomiChessDK4 : 2438 30 30 420 56.7 % 2391 21.4 %
8 RomiChessP3j : 2423 40 40 248 52.6 % 2404 14.9 %
9 RomiChessDK2 : 2418 33 33 360 53.6 % 2392 16.1 %
10 Phalanx Reborn : 2413 27 27 540 51.5 % 2403 15.6 %
11 Delphil 1.6c : 2402 27 27 517 49.9 % 2403 20.5 %
12 Horizon_4_3_173 : 2391 32 32 390 48.1 % 2405 14.1 %
13 Djinn 0.925x : 2378 28 28 510 46.7 % 2401 15.7 %
14 Horizon_4_3_175 : 2371 33 33 360 46.0 % 2399 18.1 %
15 NanoSzachy 2.7 : 2356 28 28 510 42.9 % 2405 16.1 %
16 Horizon 4.3 : 2356 30 30 420 43.1 % 2404 18.6 %
17 Zeus 1.28 : 2355 27 27 510 42.7 % 2405 19.6 %
18 GreKo 5.2 : 2279 28 29 540 31.9 % 2411 16.3 %
19 Lime 6.3 : 2166 34 35 540 19.1 % 2417 10.0 %
Re: A dangerous combination
Posted: Mon Apr 02, 2007 3:06 pm
by Dr.Wael Deeb
Great Tony,keep up the good work brother