Komodo 4.1

[lkaufman]

Larry, the Houdini improvements are benchmarked against 9 different engines. Within the confidence interval of the different tests the Elo differences I quoted above are also verified against these 9 engines.

Robert

I can't think of 9 engines strong enough to use as test opponents for Houdini, unless you are using various Ippo clones which in the case of Houdini would amount to self-testing. Even Rybka would be somewhat questionable as an opponent due to the similarity. You could use weaker engines if you give them large time handicaps, but that is of course very inefficient.

[Vinvin]

Larry, the Houdini improvements are benchmarked against 9 different engines. Within the confidence interval of the different tests the Elo differences I quoted above are also verified against these 9 engines.

Robert

I can't think of 9 engines strong enough to use as test opponents for Houdini, unless you are using various Ippo clones which in the case of Houdini would amount to self-testing. Even Rybka would be somewhat questionable as an opponent due to the similarity. You could use weaker engines if you give them large time handicaps, but that is of course very inefficient.

I'd bet :
Houdini 2.0c
Houdini 1.5a
Stockfish 2.2.2
Critter 1.2
Komodo 4
Strelka 5.5
Naum 4.2
Chiron 1.1a
Fritz 13

The latest is 187 Elo below the best Houdini.

[lkaufman]

Larry, the Houdini improvements are benchmarked against 9 different engines. Within the confidence interval of the different tests the Elo differences I quoted above are also verified against these 9 engines.

Robert

I can't think of 9 engines strong enough to use as test opponents for Houdini, unless you are using various Ippo clones which in the case of Houdini would amount to self-testing. Even Rybka would be somewhat questionable as an opponent due to the similarity. You could use weaker engines if you give them large time handicaps, but that is of course very inefficient.

I'd bet :
Houdini 2.0c
Houdini 1.5a
Stockfish 2.2.2
Critter 1.2
Komodo 4
Strelka 5.5
Naum 4.2
Chiron 1.1a
Fritz 13

The latest is 187 Elo below the best Houdini.

Well, H2, H1.5, and Strelka are basically one engine, while the bottom three are really a bit too weak to make good test opponents (the margin of error is larger when the elo difference is large), though of course he could be using them. I think you forgot Rybka 4. Even Critter is a bit too similar to Houdini to make a good test opponent, though that's a matter of opinion. Really, only Komodo and Stockfish are good opponents for Houdini to test against.

[Uri Blass]

Where are the parameters for all these versions (2z, 2s, 2s2) posted? I could probably estimate whether the changes were substantial enough to account for a measurable elo gain. From the discussion it sounds like there were time control changes, which might have favored 5 + 3 over 2 + 2.

2s2 is not public but here are the setting for other versions copied from the rybka forum

http://rybkaforum.net/cgi-bin/rybkaforu ... #pid427248



Houdini 2.0c => 5D717974702A2020202020202020202020202020202020202020202020201E2328161C1E20
Houdini 2.0Baracuda => 5F758274733021212220222020201E1E1E1E20202020202222202222221E1E2328161C1E20
Houdini 2.0T3 => 5D717974702A2020202020202020202020202020202020202020202020201E23281628323C
Houdini 2.0Z => 5A7179746E2A2020202020202020202020202020202020202020202020201E22281628323C

Houdini 2.0S => 5A71797469372020202020202020202020202020202020202020202020201E23281628323C

[melajara]

Well Larry, it seems you have trouble admitting Houdini 3 DEV is actually circa 50 ELO stronger than Houdini 2c and I can imagine why, it makes very implausible to have Komodo 5 catching up with the new Houdini

Or maybe it is Houdart's ploy to entice you to get out of the wood and release, at least, Komodo 4.1 to use the rapidly closing marketing window of opportunity to offer the TOP engine in the world!

I think such tactics to hide and carefully wait for the opposition to declare its next move is negative for the community at large. I explained it in a former post which got ignored.


Mr Houdart, Dailey, Vida and other chess engine authors just behind this trio, IMHO you should publish updates to your engines as soon as having say 20 ELO improvement and not delay.

Of course it will make the competition's job easier as each top opponent would adapt and leverage on your new version, but so will you do against them for your next move!

By abiding to this gentleman agreement policy, the computer chess community would witness much more rapid progress toward what really matters: better and better chess!

[Houdini]

There are practical constraints for a commercial chess engine that is sold as DVD version in shops throughout the world, too frequent updates would be inconvenient for the dealer network.
A yearly release seems a good compromise.

[Don]

Well Larry, it seems you have trouble admitting Houdini 3 DEV is actually circa 50 ELO stronger than Houdini 2c and I can imagine why, it makes very implausible to have Komodo 5 catching up with the new Houdini

Or maybe it is Houdart's ploy to entice you to get out of the wood and release, at least, Komodo 4.1 to use the rapidly closing marketing window of opportunity to offer the TOP engine in the world!

I think such tactics to hide and carefully wait for the opposition to declare its next move is negative for the community at large. I explained it in a former post which got ignored.


Mr Houdart, Dailey, Vida and other chess engine authors just behind this trio, IMHO you should publish updates to your engines as soon as having say 20 ELO improvement and not delay.

Of course it will make the competition's job easier as each top opponent would adapt and leverage on your new version, but so will you do against them for your next move!

By abiding to this gentleman agreement policy, the computer chess community would witness much more rapid progress toward what really matters: better and better chess!

I think you are way too paranoid if you imagine that is what is going on here. It's foolish to worry about stuff like that because there is nothing you can do about someone else's progress, you just have to worry about your own progress and you have no control over what everyone else does. We are not trying to "time" our releases, we release when it's not too often and when we have something worthy of release.

[MM]

For your info, at the same 2'+2" TC the current Houdini 3 DEV version scores 57.2% against Houdini 2.0 (+50 +/- 4 Elo), and 60.2% against Houdini 1.5 (+72 +/- 4 Elo). Again using 9,000 game test matches.

Looking forward to Komodo 5.

Cheers,
Robert

Hi Robert,

thank you for the info, please don't stop development, you still have more than 1 month, possibly you could reach the monster percentage of 60% against H2.0, could it be your goal?

Best Regards

[MM]

Well Larry, it seems you have trouble admitting Houdini 3 DEV is actually circa 50 ELO stronger than Houdini 2c and I can imagine why, it makes very implausible to have Komodo 5 catching up with the new Houdini

Or maybe it is Houdart's ploy to entice you to get out of the wood and release, at least, Komodo 4.1 to use the rapidly closing marketing window of opportunity to offer the TOP engine in the world!

I think such tactics to hide and carefully wait for the opposition to declare its next move is negative for the community at large. I explained it in a former post which got ignored.


Mr Houdart, Dailey, Vida and other chess engine authors just behind this trio, IMHO you should publish updates to your engines as soon as having say 20 ELO improvement and not delay.

Of course it will make the competition's job easier as each top opponent would adapt and leverage on your new version, but so will you do against them for your next move!

By abiding to this gentleman agreement policy, the computer chess community would witness much more rapid progress toward what really matters: better and better chess!

I think you are way too paranoid if you imagine that is what is going on here. It's foolish to worry about stuff like that because there is nothing you can do about someone else's progress, you just have to worry about your own progress and you have no control over what everyone else does. We are not trying to "time" our releases, we release when it's not too often and when we have something worthy of release.

Well honestly i think Larry seems worry about Houdini development. I could be wrong but i think that Komodo wants to overtake Houdini. But the thing that impresses me more is that Larry keeps on talking about Houdini, it seems to be an obsession.

Don, you said you have a Komodo version much stronger even than 4.1. So the question is natural. Why don't you release it? I don't ask for a MP version. But at least a SP version.

Robert Houdart has specified the improvement of H3 and announced a release date. That's what customers and fans want. If you have a good product right now, why don't you do the same thing?

I've been buying all kinds of engines for years but lately i'm getting tired to buy everything, and i'm getting more tired to wait for something that i don't even know how strong it is.


Thank you.

Best Regards

[MM]

Well Larry, it seems you have trouble admitting Houdini 3 DEV is actually circa 50 ELO stronger than Houdini 2c and I can imagine why, it makes very implausible to have Komodo 5 catching up with the new Houdini

Or maybe it is Houdart's ploy to entice you to get out of the wood and release, at least, Komodo 4.1 to use the rapidly closing marketing window of opportunity to offer the TOP engine in the world!

I think such tactics to hide and carefully wait for the opposition to declare its next move is negative for the community at large. I explained it in a former post which got ignored.


Mr Houdart, Dailey, Vida and other chess engine authors just behind this trio, IMHO you should publish updates to your engines as soon as having say 20 ELO improvement and not delay.

Of course it will make the competition's job easier as each top opponent would adapt and leverage on your new version, but so will you do against them for your next move!

By abiding to this gentleman agreement policy, the computer chess community would witness much more rapid progress toward what really matters: better and better chess!

I think the most you wrote is extremely logical and i subscribe it.

Thank you.

Best Regards