TCEC Elite Match - Season 2 (150m + 30s) LIVE! starts 12/4

[S.Taylor]

"Interesting", "amazing", "remarkable", Now at round 18, houdini is 2 up in the lead, but the end result is far from clear, and it does seem like Rybka may have removed some of its dumbing down bugs, which it was highly reputed to have been made with. This brings it so close to Houdini, if not equal, or even more, that we cannot yet be satisfied that Houdini is above Rybka. And Houdini is not on plan to have any further release before September. So Rybka has therefore not been adequately responded to.

[Jouni]

Result: 22 - 18 for Houdini means +35. So Rybka 4.1 was 25 points better than 4 = quite similar to some other tests with much faster time control.
BTW I don't understand critic for Rybka time management, isn't it always good to use all available time!?

Jouni

[Martin Thoresen]

Final standings:

Code: Select all

N Engine         Pts    SB     Houdini 1.5a  Rybka 4.1    

1 Houdini 1.5a  22.0  396.00                 + 9 = 26 - 5 
2 Rybka 4.1     18.0  396.00   + 5 = 26 - 9              

After an exciting start where Rybka took an early lead Houdini
dominated the later part of the match and successfully defended
its TCEC Champion title.

See you all in season 3!

[S.Taylor]

Final standings:

Code: Select all

N Engine         Pts    SB     Houdini 1.5a  Rybka 4.1    

1 Houdini 1.5a  22.0  396.00                 + 9 = 26 - 5 
2 Rybka 4.1     18.0  396.00   + 5 = 26 - 9              

After an exciting start where Rybka took an early lead Houdini
dominated the later part of the match and successfully defended
its TCEC Champion title.

See you all in season 3!

Thanks, it was very exciting, and it makes good sense now.

What changes might be done in the elite match, by Season 3?

Anyone has any knowledge or speculation?

[ernest]

Result: 22 - 18 for Houdini means +35.

Actually, Result: 22 - 18 for Houdini means +35 ±65 Elo (at 95% probability)

[Houdini]

Final standings:

Code: Select all

N Engine         Pts    SB     Houdini 1.5a  Rybka 4.1    

1 Houdini 1.5a  22.0  396.00                 + 9 = 26 - 5 
2 Rybka 4.1     18.0  396.00   + 5 = 26 - 9              

After an exciting start where Rybka took an early lead Houdini
dominated the later part of the match and successfully defended
its TCEC Champion title.

See you all in season 3!

Thanks, Martin!
I've been without internet for the last 14 days and have completely missed the Elite Match, but it appears to have been a thrilling event and I'll certainly replay the decisive games.

The 22-18 result (H1.5 v R4.1) is not statistically different from the previous 23.5-16.5 result (H1.5 v R4).
If one combines the result into a global 45.5-34.5 match result, it simply confirms that H1.5 is about 50 Elo stronger than R4/R4.1.

Looking forward to TCEC Season 3!

Robert

[beram]

EC Champion title.

I've been without internet for the last 14 days and have completely missed the Elite Match, but it appears to have been a thrilling event and I'll certainly replay the decisive games.

The 22-18 result (H1.5 v R4.1) is not statistically different from the previous 23.5-16.5 result (H1.5 v R4).
If one combines the result into a global 45.5-34.5 match result, it simply confirms that H1.5 is about 50 Elo stronger than R4/R4.1.

Looking forward to TCEC Season 3!

Robert

Hi Robert,

It really was interesting match. R4.1 played better than R4.
Will we see an approved Houdini in season 3 ?

grts Bram

[Houdini]

Hello Bram,

R4.1 played better than R4.

Did she? When I only look at the scores (haven't played any of the games yet), in the final 30 games of this match H1.5 beat R4.1 by the score of 18.5-11.5 (+8 -1 =21), that's an even more pronounced result than the first Elite Match with R4.

I'm always surprised that people so easily say that "R4.1 is better than R4 at long TC" when the Rybka team itself announced R4.1 as a bug fix release without impact on the strength.

Will we see an approved Houdini in season 3 ?

There will certainly be no Houdini 2.0 release before September, so probably not.

Cheers,
Robert

[beram]

Hello Bram,

R4.1 played better than R4.

Did she? When I only look at the scores (haven't played any of the games yet), in the final 30 games of this match H1.5 beat R4.1 by the score of 18.5-11.5 (+8 -1 =21), that's an even more pronounced result than the first Elite Match with R4.

I'm always surprised that people so easily say that "R4.1 is better than R4 at long TC" when the Rybka team itself announced R4.1 as a bug fix release without impact on the strength.

Will we see an approved Houdini in season 3 ?

There will certainly be no Houdini 2.0 release before September, so probably not.

Cheers,
Robert

Well Robert, my opinion is based on watching and replaying the games in the last matches in TCEC EM. And by doing so i found that Rybka 4.1 really played better than Rybka 4 in last match. I have made some comments up to game 18 - I will share them with you all, when I am done with all 40 games.

I am already very glad that there will come an improved Houdini 2.0 sooner or later.

Cheers Bram

[Martin Thoresen]

Robert,

Yes we were wondering where you went after you stopped by in the final round of TCEC A.

I agree with Bram though, it seems Rybka 4.1 played "better" than 4.0 even though the scores in the two EMs are somewhat the same.

Best,
Martin