Chess programs are worth more than ever! get all?

[Dann Corbit]

Not only that, but I bet that top freeware programs are well above the legendary deep blue as well.

Not to be contrary, but I doubt it.

I'm pretty sure Rybka 1.0 Beta could beat Deep Blue, Rybka Beta was just far ahead of its time and revolutionized computer chess, without it, I think we would be still stuck getting marginal results (Like we were for some time).

Maybe Rybka 1.0 Beta would need a 64bit processor, maybe it would win by a small margin, but I just can't imagine Deep Blue beating it.

Tragically, it's another "How many angels can dance on the head of a pin?" experiment.

[Mark]

Not only that, but I bet that top freeware programs are well above the legendary deep blue as well.

Not to be contrary, but I doubt it.

I'm pretty sure Rybka 1.0 Beta could beat Deep Blue, Rybka Beta was just far ahead of its time and revolutionized computer chess, without it, I think we would be still stuck getting marginal results (Like we were for some time).

Maybe Rybka 1.0 Beta would need a 64bit processor, maybe it would win by a small margin, but I just can't imagine Deep Blue beating it.

Tragically, it's another "How many angels can dance on the head of a pin?" experiment.

Maybe we could make an educated guess. I'm sure someone has analyzed all of the Deep Blue - Kasparov games by now with Rybka. Were any Deep Blue errors found or improvements in play?

Regards,

Mark

[Dann Corbit]

Not only that, but I bet that top freeware programs are well above the legendary deep blue as well.

Not to be contrary, but I doubt it.

I'm pretty sure Rybka 1.0 Beta could beat Deep Blue, Rybka Beta was just far ahead of its time and revolutionized computer chess, without it, I think we would be still stuck getting marginal results (Like we were for some time).

Maybe Rybka 1.0 Beta would need a 64bit processor, maybe it would win by a small margin, but I just can't imagine Deep Blue beating it.

Tragically, it's another "How many angels can dance on the head of a pin?" experiment.

Maybe we could make an educated guess. I'm sure someone has analyzed all of the Deep Blue - Kasparov games by now with Rybka. Were any Deep Blue errors found or improvements in play?

Regards,

Mark

I suspect that if we took Rybka on 8 CPUs verses Rybka on 8 CPUs and ran 7 games at correspondence time control, and then analyzed the games carefully after the fact, we would find mistakes (after all, one program is going to lose some of the time). But what would that tell us?

I think the only way to know how Deep Blue would do is run SSDF or CCRL or CEGT style battles until we have 1000 games and then we would know exactly the real truth of the matter. Without that, I see only speculation. Sure, we know Deep Blue was 'Real Strong'...
But exactly how strong cannot be determined from 7 games (the last machine was stronger than the first).

[S.Taylor]

I still agree with what I used to say years ago, that I strongly believe there exists a way to determine things from a small sample of games.
Highly rated players are often saying what they seem to be claiming is proffesional judgement on these things. So maybe they sometimes DO know what they're are talking about. You, Dan, may be very scientifically minded and realistic and very good at chess, but some GM's might outweigh all that when they give educated opinions on chess programs.
And even if not, I personally believe there has to be a way of studying it from a small sample of games, even if it hasn't ever been officially formulated.

And anyway, what are you talking about? getting DB put back together again for 1000 games? Sure, that would be interesting!

But didn't DB once fail to Fritz 3? Even if this was just one unlucky happening, it's enough to show that DB was surely below Rybkas level.

[Dann Corbit]

I still agree with what I used to say years ago, that I strongly believe there exists a way to determine things from a small sample of games.
Highly rated players are often saying what they seem to be claiming is proffesional judgement on these things. So maybe they sometimes DO know what they're are talking about. You, Dan, may be very scientifically minded and realistic and very good at chess, but some GM's might outweigh all that when they give educated opinions on chess programs.
And even if not, I personally believe there has to be a way of studying it from a small sample of games, even if it hasn't ever been officially formulated.

You can be right, and yet we would never know it without math and science. You are talking about feelings and I am talking about proof. The feelings can be right, but they are not certain util proven with formal proof.

To say it another way. You 'might' be right. But your intuition or the inutition of 1000 GMs cannot prove the matter. Even if they made the same supposition as you and even if they were correct we would not *know* that they are correct. We would only suspect it.

And anyway, what are you talking about? getting DB put back together again for 1000 games? Sure, that would be interesting!

It won't happen, though I would love to see it.

But didn't DB once fail to Fritz 3? Even if this was just one unlucky happening, it's enough to show that DB was surely below Rybkas level.

In 1995 Fritz beat a single chip prototype of the kind that was eventually developed into the chess processors used by Deep Blue. It was perhaps 1% as strong as Deep Blue.

[S.Taylor]

In 1995 Fritz beat a single chip prototype of the kind that was eventually developed into the chess processors used by Deep Blue. It was perhaps 1% as strong as Deep Blue.

Oh!
And are you SURE about that? I had thought it was after the 1996 match, and WITH most of its strength.
If as you say, then that's REALLY meaningless.
(But I think 1% would be a little underestimation )

[Dann Corbit]

In 1995 Fritz beat a single chip prototype of the kind that was eventually developed into the chess processors used by Deep Blue. It was perhaps 1% as strong as Deep Blue.

Oh!
And are you SURE about that? I had thought it was after the 1996 match, and WITH most of its strength.
If as you say, then that's REALLY meaningless.
(But I think 1% would be a little underestimation )

The actual machine had chips that were faster than the single chip used in the WC match of 1995, and there were 480 of them used in Deeper Blue (IIRC). 1% implies that the SMP loss was about 80%. I think that is a conservative estimate.

Of course, the modern engines might be stronger. But we have no *provable* way of knowing.

[Jim Walker]

Not to be contrary but....
I think the revolution started with Fruit. Rybka just put into overdrive.
JMHO
Jim