100 long games Rybka 4 vs Houdini 1.03a

[Milos]

Well, I don't do any beta testing for any engine.

Not true. It is well known that you ppl in CCRL did beta test Rybka 4, it was even part of your list for some time. Now you will probably say you were not part of it, won't you?

How many Rybka testing results do you see me publishing? I'm a fan of all the chess engines that I test. Not sure why you persist in calling me a Rybka fan, just for not recognizing or testing the Ippos.

Lets see...
You feel the urge to defend Rybka results any time they seem to be negative.
You had a personal email correspondence with Vas regarding the issue of Ippos, or at least you were dealing as Vas's proxy regarding this topic.
You stubbornly reject any argument and proof regarding Rybka being a clone of Fruit, while you insist on Ippos being clones despite much less evidence (well that speaks more about your hypocrisy).
You made a huge effort as a moderator to protect Vas's business interest.
Finally, you never ever made a slightest critic of Vas's behavior that even some other Rybka fans admit.

Of course I like to occasionally watch some of the engine v engine games that I run. Can't see anything unhealthy in that.

There is nothing bad about occasionally watching. Doing only that, or thinking you are becoming smarter by watching is what is unhealthy.

Lastly, your manners stink.

Says someone who likes to put pictures of feces to others ppl work. Well this almost comes as a compliment coming from somebody like you .

[tomgdrums]

Well, I don't do any beta testing for any engine.

Not true. It is well known that you ppl in CCRL did beta test Rybka 4, it was even part of your list for some time. Now you will probably say you were not part of it, won't you?

How many Rybka testing results do you see me publishing? I'm a fan of all the chess engines that I test. Not sure why you persist in calling me a Rybka fan, just for not recognizing or testing the Ippos.

Lets see...
You feel the urge to defend Rybka results any time they seem to be negative.
You had a personal email correspondence with Vas regarding the issue of Ippos, or at least you were dealing as Vas's proxy regarding this topic.
You stubbornly reject any argument and proof regarding Rybka being a clone of Fruit, while you insist on Ippos being clones despite much less evidence (well that speaks more about your hypocrisy).
You made a huge effort as a moderator to protect Vas's business interest.
Finally, you never ever made a slightest critic of Vas's behavior that even some other Rybka fans admit.

Of course I like to occasionally watch some of the engine v engine games that I run. Can't see anything unhealthy in that.

There is nothing bad about occasionally watching. Doing only that, or thinking you are becoming smarter by watching is what is unhealthy.

Lastly, your manners stink.

Says someone who likes to put pictures of feces to others ppl work. Well this almost comes as a compliment coming from somebody like you .

So if it is unhealthy to think i am becoming smarter by watching engine games, does it also mean that it is bad to think I am becoming dumber by reading your posts?

I think Glenn Beck owns Rybka stock.

[Uri Blass]

As for asking for the pgn file ... you might not know this but quite a few people here actually do go through some of their engine games and ENJOY IT!

Sure you might think you must be really smart since you are watching a game of two engines 1000+ elo stronger than you. Better than watching a real GM's match isn't it? Oh man, get a life, don't be just like Banks, trust me, it's not healty.

It is a good way to see where an engine is strong and where it is weak. There are many ways to make use of a database, you can search certain specific end games and see how the engine plays them, you can look at the losses and try to figure out if a certain poor evaluation caused a loss, or maybe a bug that occurs occasional hits ... but of course you knew all that.

You definitely have no clue about chess engines and even less about chess programming. You are just talking BS and trolling, as usual.
If you are GM that understands computer chess well (and these you can't even count on fingers of one hand) you can improve someones engine if his engine is in 2500 elo zone.
Improving 3200 elo engine by looking for weaknesses in play, that's just ridiculous.
Why do you think Vas kicked Larry K., because of his usefulness?
Do you have any clue how many evaluation things from Larry actually came into Rybka 3 code? Real things not Larry's stories?
I'll give you a secret. None, not a single coefficient from Larry is identical in R3. All are heavily automatically tuned (if you even understand what this means) and some are more than 100% different than what Larry originally proposed. And Larry is a GM that has the best knowledge of computer chess among all GMs.

By the way, I thought a while back that in one of your tantrums, you had decided to leave this forum and go somewhere else ... what made you come back here with your tail between your legs?

I can only say you are either dreaming or using too much chemistry that makes you loose contact with reality...

I disagree with you and I believe that strong players can help to improve stockfish's evaluation based on watching stockfish's games and I guess that it is going to happen in the future(We do not have rybka's code so I will say nothing about rybka).

My point is that I disagree with
"Improving 3200 elo engine by looking for weaknesses in play, that's just ridiculous."

[Albert Silver]

My point is that I disagree with
"Improving 3200 elo engine by looking for weaknesses in play, that's just ridiculous."

It could be that he is just very weak and thus the concept looks like science fiction. There are many weaknesses in its evaluation that do not require a GM to see, but if even these are beyond one's reach, then the idea of correcting them must truly seem ludicrous.

Here is an easy example:

[D]1R6/1P4k1/7p/1r5P/6K1/8/8/8 w - - 0 1

Analysis by Deep Rybka 4 x64:

1.Kf4 Kh7 2.Ke4 Kg7 3.Kd4 Rb6 4.Kc5 Rb3 5.Kc6 Rc3+ 6.Kb5 Rb3+[] 7.Kc5 Rb1 8.Kc6 Rc1+ 9.Kb6 Rb1+[] 10.Kc7 Rc1+ 11.Kd7 Rb1 12.Kc6 Rc1+ 13.Kb6 Rb1+[] 14.Kc7 Rc1+ 15.Kd7 Rb1 16.Kc6
+- (3.35) Depth: 31 00:01:27 1359kN, tb=6850
1.Kf4 Rb3 2.Ke5 Kh7 3.Kd6[] Kg7 4.Kc7 Rc3+ 5.Kb6 Rb3+ 6.Ka5 Rb1 7.Ka6 Ra1+[] 8.Kb6[] Rb1+[] 9.Kc5 Rc1+ 10.Kb6 Rb1+ 11.Kc5 Rc1+ 12.Kb6 Rb1+ 13.Kc5 Rc1+ 14.Kb6 Rb1+ 15.Kc5 Rc1+ 16.Kb6
+- (3.35) Depth: 32 00:02:05 2225kN, tb=10450

Houdini also misevaluates and gives it +1.54, while Stockfish 1.9.1 gets it right (0.00) after a bit of thought.

[Graham Banks]

Well, I don't do any beta testing for any engine.

Not true. It is well known that you ppl in CCRL did beta test Rybka 4, it was even part of your list for some time. Now you will probably say you were not part of it, won't you?

Indeed I deny it. Shaun Brewer was the only CCRL tester who helped beta test Rybka 4.

How many Rybka testing results do you see me publishing? I'm a fan of all the chess engines that I test. Not sure why you persist in calling me a Rybka fan, just for not recognizing or testing the Ippos.

Lets see...
You feel the urge to defend Rybka results any time they seem to be negative.
You had a personal email correspondence with Vas regarding the issue of Ippos, or at least you were dealing as Vas's proxy regarding this topic.
You stubbornly reject any argument and proof regarding Rybka being a clone of Fruit, while you insist on Ippos being clones despite much less evidence (well that speaks more about your hypocrisy).
You made a huge effort as a moderator to protect Vas's business interest.
Finally, you never ever made a slightest critic of Vas's behavior that even some other Rybka fans admit.

More than once I have stated that Vas should have said more regarding the Ippo issue.

Of course I like to occasionally watch some of the engine v engine games that I run. Can't see anything unhealthy in that.

There is nothing bad about occasionally watching. Doing only that, or thinking you are becoming smarter by watching is what is unhealthy.

I have a job, a family and other interests. I'd like to be able to watch more games than I do though.

Lastly, your manners stink.

Says someone who likes to put pictures of feces to others ppl work. Well this almost comes as a compliment coming from somebody like you .

Now now - I quite enjoyed those logos.

[Milos]

My point is that I disagree with
"Improving 3200 elo engine by looking for weaknesses in play, that's just ridiculous."

It could be that he is just very weak and thus the concept looks like science fiction. There are many weaknesses in its evaluation that do not require a GM to see, but if even these are beyond one's reach, then the idea of correcting them must truly seem ludicrous.

Neither Uri nor you do not understand.
The point is not improving strong engine's evaluation. The point is improving its overall strength. There is quite a bit of difference between these two.
I would not expect from you to understand this, however, I'm surprised that Uri doesn't...

[Uri Blass]

My point is that I disagree with
"Improving 3200 elo engine by looking for weaknesses in play, that's just ridiculous."

It could be that he is just very weak and thus the concept looks like science fiction. There are many weaknesses in its evaluation that do not require a GM to see, but if even these are beyond one's reach, then the idea of correcting them must truly seem ludicrous.

Neither Uri nor you do not understand.
The point is not improving strong engine's evaluation. The point is improving its overall strength. There is quite a bit of difference between these two.
I would not expect from you to understand this, however, I'm surprised that Uri doesn't...

If you improve the evaluation without the price of doing the program significantly slower then you also improve the strength of the program(and I believe that it is easy to do it at least for stockfish).

I suggested some change in parameters for mobility evaluation in stockfish(and I remember that someone found that testing at fast time control suggested 2 elo improvement that is below the statistical error but a reason to believe that my idea is productive)

I do not know if people did more testing of my idea(if testing suggest small improvement then maybe bigger change to the same direction can give even more improvement).

I do not claim that testing is useless but only that chess players can be productive to improve the evaluation even if they only suggest to change parameters(numbers inside the program) because there are many parameters that you can try to change and having strong player that guess the direction to change is going to give faster improvement than testing a random change.

Here is a link for the discussion about the change that I suggested

http://talkchess.com/forum/viewtopic.ph ... 7&start=10

[Ralph Stoesser]

If it's easy for you to improve SF, then please do it. Push SF to the pole position.

[mcostalba]

I disagree with you and I believe that strong players can help to improve stockfish's evaluation based on watching stockfish's games

Here is my opinion regarding this point.

Implementation of any evaluation idea in chess engines is always the combined effect of two contributions.

1 - The actual idea, for instance a new way to evaluate pawns structure or a new way to evaluate passed pawns.

2 - The tuning of the coefficents that _weights_ the idea among the other evaluation terms.


IMHO a strong player could be effective regarding point one, i.e. to propose an interesting idea. Then the next step is "to make the idea to work" (because normally even a good idea does _not_ work at first try) finding the right coefficents bounded to the idea in the evaluation code and this is a task up to real games testing possibly without human intervention.

The wrong thing to do IMHO is to pretend a human is effective in both the first and the second points.

Another wrong thing to do is to consider an human useless because is not able to come up with point 2.

[Tord Romstad]

Houdini also misevaluates and gives it +1.54, while Stockfish 1.9.1 gets it right (0.00) after a bit of thought.

It does, but it gets it right by search, not by evaluation. SF misevaluates this position just as badly as the two other programs mentioned.