LS-ratinglist (news & comments)

[Vinvin]

The result of Stockfish DD with adaptive Contempt (by Jörg Oster) is now online.

http://ls-ratinglist.beepworld.de

Stefan

Copying the interesting note from the page :

Stockfish DD adapC =Stockfish DD with adaptive Contempt by Jörg Oster (download source and sse42-compile). The real interesting fact of this Stockfish-mod is, that the ratio of draw by 3fold-repetition is only 17% (Stockfish DD default: 28% (!!!), which means around 40% less draws by 3fold-repetition for the adapted Contempt-version), so this mod has sucessfully eliminated a real weakness of the Stockfish engine (too many draws by 3fold-repetition), congrats to Jörg Oster. Next test: Komodo TCEC

Personal note : the draw rate is similar anyway : 47% to 46%

Code: Select all

5 Stockfish DD x64s        3120    5    5 11000   57%  3069   47% 
6 Stockfish DD adapC x64s  3119    5    5 10000   57%  3069   46%

[ouachita]

I'm wondering why so little mention is made here about the strong performance of Robodini 1.1 x64, stronger than SF

[ThomasJMiller]

I'm wondering why so little mention is made here about the strong performance of Robodini 1.1 x64, stronger than SF

Robodini is a 100% clone of Houdini 3, so no surprise.

[pohl4711]

I'm wondering why so little mention is made here about the strong performance of Robodini 1.1 x64, stronger than SF

Robodini is a 100% clone of Houdini 3, so no surprise.

No, Robodini is a Houdini 3 core in a Robbolito frame. So it is not a 100%-clone. And the underpromotion bug of Robodini is not a part of Houdini. So Robodini is near at 100% -clone. But not exactly 100%, so I had to test it, following my test-conditions (testing all engines, which are not exactly 100%-clones).

Stefan

[ThomasJMiller]

I'm wondering why so little mention is made here about the strong performance of Robodini 1.1 x64, stronger than SF

Robodini is a 100% clone of Houdini 3, so no surprise.

No, Robodini is a Houdini 3 core in a Robbolito frame. So it is not a 100%-clone. And the underpromotion bug of Robodini is not a part of Houdini. So Robodini is near at 100% -clone. But not exactly 100%, so I had to test it, following my test-conditions (testing all engines, which are not exactly 100%-clones).

Stefan

Since it seems you like exact figures I will give you: Robodini 1.1 has a similarity of 99.27% to Houdini 3 according to sim03 and my testing conditions.

[pohl4711]

Since it seems you like exact figures I will give you: Robodini 1.1 has a similarity of 99.27% to Houdini 3 according to sim03 and my testing conditions.

I do not doubt this. And I really dont like engines like Robodini, and I dont understand R.Vida creating this one. But as I said: Not 100%-clone, so I had to do a testrun - so say my test-conditions (and if I do/allow exceptions from my own test-conditions, I can throw them away)...But definitly nobody needs Robodini...

Stefan

[ouachita]

I dont understand R.Vida creating this one.

spite?

[IGarcia]

I dont understand R.Vida creating this one.

spite?

It was a bet on how fast a engine can be RE.
A very shor ressumme of events as I understand:

Houdart said his engine houdini 2 was RE in just 1 and half years (or something like that). Then Vida told that was true if the RE process had started on engine release. As for Vida, the RE process was much faster.

Later someone made a bet, and he in few days (weeks) come with that engine, a RE houdini 3, and winning the bet.

The problem behind that fast RE is to be accomplished you have to have a source frame, witch was robolito. This makes many to ask how original is houdini...

To be complete fair, Houdart has told about Vida access to houdini code, so, is probably the source frame is not robolito but an previous version of houdini.

So, in the end will be hard to know what is true and what is not, because there is a lot of interest, pride, vanity, spite, threats, money and lies.

Regards

[pohl4711]

The result of Stockfish 131210 is now online.

http://ls-ratinglist.beepworld.de


Stefan

(Perhaps you have to clear your browsercache or reload the website)

[velmarin]

I dont understand R.Vida creating this one.

spite?

It was a bet on how fast a engine can be RE.
A very shor ressumme of events as I understand:

Houdart said his engine houdini 2 was RE in just 1 and half years (or something like that). Then Vida told that was true if the RE process had started on engine release. As for Vida, the RE process was much faster.

Later someone made a bet, and he in few days (weeks) come with that engine, a RE houdini 3, and winning the bet.

The problem behind that fast RE is to be accomplished you have to have a source frame, witch was robolito. This makes many to ask how original is houdini...

To be complete fair, Houdart has told about Vida access to houdini code, so, is probably the source frame is not robolito but an previous version of houdini.

So, in the end will be hard to know what is true and what is not, because there is a lot of interest, pride, vanity, spite, threats, money and lies.

Regards

Although it's easier to explain it this way,
the product was never released as a motor, the product is not support,

 and asked the author of the product could not be proved.
We need more not to try.