Trying to improve lazy smp

[Sven]

Hi.
I have done a self test at 25+0.03 with 4 threads:

Code: Select all

Rank Name            Elo    +    - games score oppo. draw
1 Andscacs 0.74039    59    7    6  2512   69%   -59   45%
2 Andscacs 0.73174   -59    6    7  2512   31%    59   45%

The difference in % is equivalent to 140 elo, but bayeselo shows +-59. Someone knows what I'm interpreting bad? I used bayeselo for more than a year and I did not compared the % until now with standard elo.

The difference between 59 and -59 is 118 which is quite a good match for 140.

Yes, but with standar elo, 69% is 140 elo, not 118. Is simply different for bayeselo?

I think the draw rate plays a role but others can explain that better than me.

[AlvaroBegue]

Hi.
I have done a self test at 25+0.03 with 4 threads:

Code: Select all

Rank Name            Elo    +    - games score oppo. draw
1 Andscacs 0.74039    59    7    6  2512   69%   -59   45%
2 Andscacs 0.73174   -59    6    7  2512   31%    59   45%

The difference in % is equivalent to 140 elo, but bayeselo shows +-59. Someone knows what I'm interpreting bad? I used bayeselo for more than a year and I did not compared the % until now with standard elo.

The difference between 59 and -59 is 118 which is quite a good match for 140.

Yes, but with standar elo, 69% is 140 elo, not 118. Is simply different for bayeselo?

I think the draw rate plays a role but others can explain that better than me.

Yes, the answer is somewhere in this page: http://remi.coulom.free.fr/Bayesian-Elo/

[...]In order to perform this calculation, it is necessary to assume a little more than the usual ELO formula. The expected score as a function of the Elo difference is not enough. We need the probability of a win, a draw and a loss as a function of the Elo difference.

[bob]

Hi.
I have done a self test at 25+0.03 with 4 threads:

Code: Select all

Rank Name            Elo    +    - games score oppo. draw
1 Andscacs 0.74039    59    7    6  2512   69%   -59   45%
2 Andscacs 0.73174   -59    6    7  2512   31%    59   45%

The difference in % is equivalent to 140 elo, but bayeselo shows +-59. Someone knows what I'm interpreting bad? I used bayeselo for more than a year and I did not compared the % until now with standard elo.

The difference between 59 and -59 is 118 which is quite a good match for 140.

Yes, but with standar elo, 69% is 140 elo, not 118. Is simply different for bayeselo?

What draw rate do you assume for your 69 = 140?

[cdani]

What draw rate do you assume for your 69 = 140?

I just taked this 140 from FIDE (in fact 141):

https://www.fide.com/fide/handbook.html ... ew=article

[cdani]

Yes, the answer is somewhere in this page: http://remi.coulom.free.fr/Bayesian-Elo/

[...]In order to perform this calculation, it is necessary to assume a little more than the usual ELO formula. The expected score as a function of the Elo difference is not enough. We need the probability of a win, a draw and a loss as a function of the Elo difference.

Thanks!!

[bob]

What draw rate do you assume for your 69 = 140?

I just taked this 140 from FIDE (in fact 141):

https://www.fide.com/fide/handbook.html ... ew=article

That has an assumed draw rate however, which might not be correct for your test games...

[Sven]

What draw rate do you assume for your 69 = 140?

I just taked this 140 from FIDE (in fact 141):

https://www.fide.com/fide/handbook.html ... ew=article

That has an assumed draw rate however, which might not be correct for your test games...

No, the draw rate is simply not part of the underlying theoretical model of the Elo rating system as it is used by the FIDE. Bayeselo uses a different model that takes the probabilities for win, draw, and loss into account as well as other factors like colors. The table in the FIDE handbook lists percentage expectancy values mapped to rating differences (and vice versa) based on the normal distribution, no draw rate is involved there.

[Laskos]

I have been also able to do a limited test at 20+.03 with 16 cores:

Code: Select all

 1 Andscacs 0.74042      114   19   19   704   69%   -23   27%
 2 Komodo 5.1r2 64-bit    86   40   40   140   46%   114   34%
 3 Senpai 1.0              6   41   42   140   34%   114   29%
 4 Gull 1.2 x64          -40   41   43   142   27%   114   31%
 5 Hannibal 1.3x64       -66   44   46   142   26%   114   20%
 6 Protector 1.7.0       -99   45   48   140   22%   114   21%

The result is very good, clearly over 3100 elo.

I am confused, is this 16 core Andscacs against 1 core other engines? To be ahead of Komodo 5.1 is a feat, that's why I am asking.

[cdani]

I have been also able to do a limited test at 20+.03 with 16 cores:

Code: Select all

 1 Andscacs 0.74042      114   19   19   704   69%   -23   27%
 2 Komodo 5.1r2 64-bit    86   40   40   140   46%   114   34%
 3 Senpai 1.0              6   41   42   140   34%   114   29%
 4 Gull 1.2 x64          -40   41   43   142   27%   114   31%
 5 Hannibal 1.3x64       -66   44   46   142   26%   114   20%
 6 Protector 1.7.0       -99   45   48   140   22%   114   21%

The result is very good, clearly over 3100 elo.

I am confused, is this 16 core Andscacs against 1 core other engines? To be ahead of Komodo 5.1 is a feat, that's why I am asking.

Yes, 16 cores against 1 core
Is how I have done most of my tests.

[Laskos]

I have been also able to do a limited test at 20+.03 with 16 cores:

Code: Select all

 1 Andscacs 0.74042      114   19   19   704   69%   -23   27%
 2 Komodo 5.1r2 64-bit    86   40   40   140   46%   114   34%
 3 Senpai 1.0              6   41   42   140   34%   114   29%
 4 Gull 1.2 x64          -40   41   43   142   27%   114   31%
 5 Hannibal 1.3x64       -66   44   46   142   26%   114   20%
 6 Protector 1.7.0       -99   45   48   140   22%   114   21%

The result is very good, clearly over 3100 elo.

I am confused, is this 16 core Andscacs against 1 core other engines? To be ahead of Komodo 5.1 is a feat, that's why I am asking.

Yes, 16 cores against 1 core
Is how I have done most of my tests.

Good result. On equal hardware, on a par with Stockfish 2.1, Komodo 2.03, Rybka 3, stronger than the new wave of engines like Texel, Hannibal, Nirvana, Cheng.