Stockfish test on 2 cores with PONDER ON is running

[Jouni]

As I said Stockfish scales better! Have You set zuzwang detection ON? Otherwise there is no reason to use 1.7.1 version That could give needed 2 points to beat Rybka...

Jouni

[IWB]

Hi

As I said Stockfish scales better! Have You set zuzwang detection ON? Otherwise there is no reason to use 1.7.1 version That could give needed 2 points to beat Rybka...

Jouni

I have some more knowledge ... as I know the result of SF1.7.1 vs DF12 both on 2 Threads. SF has to gain much more than 1 or 2 Elo over R3 to close that gap!

I will play a more games after that 1400 to see if it will change a few elo. Nevertheless you can conclude that R3 and SF1.7.1 are as close as possible. These two enignes are definately at the same league!

Bye
Ingo

[mcostalba]

Nevertheless you can conclude that R3 and SF1.7.1 are as close as possible. These two enignes are definately at the same league!

Probably SF will end up behind R3, but for me to see that the two threads version is going to repeat the result of the single threaded one: defeating corresponding R3 version in the direct match (although only 100 games) it is a great joy

Thanks for your testing Ingo !

[Uri Blass]

You need to compare result against the same opponents.



We have the following final results:

1)Stockfish 1.7.1 JA 2T - Rybka 3 mp 2T 52.5-47.5

comparing score against common opponents:

2)Rybka3 mp 2T-Stockfish1.71 2T 59.5-59(Naum4.2 2T)
3)Rybka3 mp 2T-Stockfish1.71 2T 70-64.5(shredder12 2T)

4)Rybka3 mp 2T-Stockfish1.71 2T 68.5-68-(Shredder12 1T)
5)Rybka3 mp 2T-Stockfish1.71 2T 72.5-75-(Zappa Mexico2 2T)
6)Rybka3 mp 2T-Stockfish1.71 2T 82-80(Zappa Mexico2 1T)
7)Rybka3 mp 2T-Stockfish1.71 2T 83-85.5-(Onno 1-1-1)
8)Rybka3 mp 2T-Stockfish1.71 2T 86.5-81 (spark0.3 VC)
9)Rybka3 mp 2T-Stockfish1.71 2T 83-86(Deep Sjeng2008)

total points against same opponents(including direct match):

Rybka3 mp 2T:
47.5+59.5+70+68.5+72.5+82+83+86.5+83=652.5

Stockfish 1.7.1JA 2T:
52.5+59+64.5+68+75+80+85.5+81+86=651.5

[IWB]

Too early Uri,

I have to replay one game which was won originaly for Stockfish 1.7.1 2T vs ZM2 2T. I have a result but no PGN ... for whatevver reason. (Happens sometimes, in this case 1 out of 1300, and Zappa is a main canditat for that)

This time ZM2 2T is winning, so your data is not 1 point difference but 2.

Nevertheless all the other enigneswill go into the BayesElo as well and there are 100 games vs DF12 2T which you can include in your comparison ... in a few minutes.
Unfortunately I am stopped to make the comparision by answering to posts in threads ...

Bye
Ingo