My observations on Kasparov vs Deep Blue rematch

[sje]

My observations on Kasparov vs Deep Blue rematch:

1) Kasparov is an adult, is well experienced, and had already played in one match with Deep Blue. And he had access to advisers, although it's not known if he took any advice. Anyway, he knew the conditions of the rematch and had little right to complain as the Deep Blue team followed the rules.

2) The Deep Blue team was rather unsportsmanlike to the extent that it refused to make any of the games available that were played by the new version of the hardware. This aspect of the match, far more than any other, compromised both Kasparov's chances and also the overall quality of the event. It any other high level match, each opponent has access to hundreds, sometimes thousands, of game scores played by the other party. Deep Blue certainly had all of Kasparov's prior game scores; Kasparov was foolish to accept the match conditions that kept Deep Blue's games secret. Furthermore, IBM did much in its marketing scheme to present the match as a symmetric competition, and this was unnecessarily misleading to the public because there was non-symmetric secrecy involved.

3) Kasparov's baseless innuendo and outright charges of cheating were behavior unacceptable by any standard. In a way, this was even worse than some of Fischer's tirades about Soviet cheating as at least he was correct a few times.

4) The log files produced by the Deep Blue team were poorly formatted and could have been far more informative. I was surprised at how sloppy some of them were, like the entire logging process was a design afterthought. Well formatted and informative logs made available in real time to observers would have properly eliminated much of the controversy. Also, such logs would have been a good contribution to the historical record for other computer chess researchers. Again, we see the unfounded desire for secrecy on the part of the Deep Blue team.

5) The event did more to kill funding and support for computer chess research than any other. The prize money disbursed could have paid for years of projects with chances of good results that would have helped everyone in the field. But instead we have little more than a bunch of old circuit boards in the Smithsonian.

[bob]

My observations on Kasparov vs Deep Blue rematch:

1) Kasparov is an adult, is well experienced, and had already played in one match with Deep Blue. And he had access to advisers, although it's not known if he took any advice. Anyway, he knew the conditions of the rematch and had little right to complain as the Deep Blue team followed the rules.

2) The Deep Blue team was rather unsportsmanlike to the extent that it refused to make any of the games available that were played by the new version of the hardware. This aspect of the match, far more than any other, compromised both Kasparov's chances and also the overall quality of the event. It any other high level match, each opponent has access to hundreds, sometimes thousands, of game scores played by the other party. Deep Blue certainly had all of Kasparov's prior game scores; Kasparov was foolish to accept the match conditions that kept Deep Blue's games secret. Furthermore, IBM did much in its marketing scheme to present the match as a symmetric competition, and this was unnecessarily misleading to the public because there was non-symmetric secrecy involved.

3) Kasparov's baseless innuendo and outright charges of cheating were behavior unacceptable by any standard. In a way, this was even worse than some of Fischer's tirades about Soviet cheating as at least he was correct a few times.

4) The log files produced by the Deep Blue team were poorly formatted and could have been far more informative. I was surprised at how sloppy some of them were, like the entire logging process was a design afterthought. Well formatted and informative logs made available in real time to observers would have properly eliminated much of the controversy. Also, such logs would have been a good contribution to the historical record for other computer chess researchers. Again, we see the unfounded desire for secrecy on the part of the Deep Blue team.

5) The event did more to kill funding and support for computer chess research than any other. The prize money disbursed could have paid for years of projects with chances of good results that would have helped everyone in the field. But instead we have little more than a bunch of old circuit boards in the Smithsonian.

I don't agree with (2) above. You do realize that the thing didn't play any games at all (new hardware) before the match. The chips were very late, and had some serious issues that Hsu had to kludge around to avoid some timimg issues. This was explained in his book... I don't have his book here at home, but I think they got it running less than a week prior to the match or something very close like that...

[sje]

I don't agree with (2) above. You do realize that the thing didn't play any games at all (new hardware) before the match. The chips were very late, and had some serious issues that Hsu had to kludge around to avoid some timimg issues. This was explained in his book... I don't have his book here at home, but I think they got it running less than a week prior to the match or something very close like that...

My understanding is that none of Deep Blue's games were made available including those played with the preliminary hardware. Kasparov should have made it a condition of the match that at least a hundred or so games should have been made available some weeks before the event.

[Rolf]

My observations on Kasparov vs Deep Blue rematch:

1) Kasparov is an adult, is well experienced, and had already played in one match with Deep Blue. And he had access to advisers, although it's not known if he took any advice. Anyway, he knew the conditions of the rematch and had little right to complain as the Deep Blue team followed the rules.

Rules? Adult behavior? You must have missed my fundamental critic. In such a show the organiser is the one who must supervise the science and good manners aspect of universal fairness, in short, the ethical rules of a civilised society. If say in the 100 meter hurdles final one party were allowed to construct deeper watered holes after every two hurdles and the opposing runner as a favorite for the win accidentaly falls into such a hole you wont accuse him of not having ordered in the contracts that such holes would be forbidden. Hyatt doesnt become tired to claim that in fact a runner must do that too. Ridiculous and odd.

2) The Deep Blue team was rather unsportsmanlike to the extent that it refused to make any of the games available that were played by the new version of the hardware. This aspect of the match, far more than any other, compromised both Kasparov's chances and also the overall quality of the event. It any other high level match, each opponent has access to hundreds, sometimes thousands, of game scores played by the other party. Deep Blue certainly had all of Kasparov's prior game scores; Kasparov was foolish to accept the match conditions that kept Deep Blue's games secret. Furthermore, IBM did much in its marketing scheme to present the match as a symmetric competition, and this was unnecessarily misleading to the public because there was non-symmetric secrecy involved.

To this argument Hyatt had also said earlier: does the team know what Kasparov had recently changed in his preparation? No, but the overall play of him is known. Also his particular belief into the magic of mysteries and superstition. The finding of my analyses of the video from 2005 was that someone like Campbell is even now completely unable to realise what the point is for a human chessplayer. Because he argued that Kasparov with his play record was uninteresting because they had created a machine that played just chess and not chess against Kasparov. Now let's assume this isnt a lie, then we come closer to the evaluation of such a match setting. Has a player any chances to understand the chess of a machine in just three games with every color? If he hasnt seen anything else of the newly made robot? Unlikely. So what is the point?
The whole event degenerates to a psycho show because the human in his desperate tries to decrypt a personality walks right into superstition when something happens like the personality switch between round 1 and 2. If he then asks for details and gets impolite responses the whole mindframe is shaken because suddenly people like him, from mankind, are against him, when formerly it was always a sort of collegial togetherness bnetween the team and him as the genial chessplayer. For such a psycho war Kasparov is the worst example among the top players. Because of his weakness, superstition. Especially for a guy like him with his preparations, here without a clue, or worse, without correct clues, the main strength of him was neutralized.

So, why bother at all? What is the point here?

The point is still that such a match setting should have possibly peoven what a machine in chess could do against the best human player and his chess. His chess but not his superstition or psychological weaknesses. So here the organizers allowed to create psychological obstacles in favor of the machine who allegedly just wanted to play chess, not even chess against Kasparov. Ridiculous. It was a psycho war to get to the Prize money. The famous playing dirty of the American sport. That is not British fairness what the Brits had invented as a basis for any sport.

3) Kasparov's baseless innuendo and outright charges of cheating were behavior unacceptable by any standard. In a way, this was even worse than some of Fischer's tirades about Soviet cheating as at least he was correct a few times.

4) The log files produced by the Deep Blue team were poorly formatted and could have been far more informative. I was surprised at how sloppy some of them were, like the entire logging process was a design afterthought. Well formatted and informative logs made available in real time to observers would have properly eliminated much of the controversy. Also, such logs would have been a good contribution to the historical record for other computer chess researchers. Again, we see the unfounded desire for secrecy on the part of the Deep Blue team.

5) The event did more to kill funding and support for computer chess research than any other. The prize money disbursed could have paid for years of projects with chances of good results that would have helped everyone in the field. But instead we have little more than a bunch of old circuit boards in the Smithsonian.

If necessary more about the other points, later.

[bob]

I don't agree with (2) above. You do realize that the thing didn't play any games at all (new hardware) before the match. The chips were very late, and had some serious issues that Hsu had to kludge around to avoid some timimg issues. This was explained in his book... I don't have his book here at home, but I think they got it running less than a week prior to the match or something very close like that...

My understanding is that none of Deep Blue's games were made available including those played with the preliminary hardware. Kasparov should have made it a condition of the match that at least a hundred or so games should have been made available some weeks before the event.

That's wrong. The original deep thought, deep blue prototype, and even deep blue 1 played many public matches, where the games were available. Only deep blue 2 had no public games available because it never played a public game (or any kind of game for that matter) prior to the kasparov match, it was put together at the last minute...

DB played matches all over the country. I watched a couple at one of the supercomputing conferences where it trounced GM Robert Byrne.. So games were available, but DB2 was a complete re-design of the chips with much more evaluation hardware (some of which was not even used for the Kasparov match, in fact) and many more chess processors than its predecessor deep blue 1, which would have made the old games somewhat useless.

Kasparov's _main_ problem was listening to Friedel, who had him convinced that sparring with the current Fritz would be very similar to playing against DB2. A completely ridiculous assertion. And that is what got Kasparov's "goat" because DB2 saw far more than fritz and he was completely unprepared, although he thought he was going to blow it out as he no doubt did to fritz.

[bob]

My observations on Kasparov vs Deep Blue rematch:

1) Kasparov is an adult, is well experienced, and had already played in one match with Deep Blue. And he had access to advisers, although it's not known if he took any advice. Anyway, he knew the conditions of the rematch and had little right to complain as the Deep Blue team followed the rules.

Rules? Adult behavior? You must have missed my fundamental critic. In such a show the organiser is the one who must supervise the science and good manners aspect of universal fairness, in short, the ethical rules of a civilised society. If say in the 100 meter hurdles final one party were allowed to construct deeper watered holes after every two hurdles and the opposing runner as a favorite for the win accidentaly falls into such a hole you wont accuse him of not having ordered in the contracts that such holes would be forbidden. Hyatt doesnt become tired to claim that in fact a runner must do that too. Ridiculous and odd.

That's ridiculous. If you sign a contract that gives your opponent the right to change the course, then if they do so, it is not _their_ fault, it is _yours_. The match conditions for 1997 were exactly the same as they were for 1996. Kasparov signed on the dotted line for both. It is somewhat like signing a mortgage to buy a house, and then a year later you finally figure out you just can't pay that much per month and are going to default on the loan and lose the house, and then you blame the bank for letting you buy a house you could not afford. You have to accept responsibility for your own actions. Kasparov didn't.

2) The Deep Blue team was rather unsportsmanlike to the extent that it refused to make any of the games available that were played by the new version of the hardware. This aspect of the match, far more than any other, compromised both Kasparov's chances and also the overall quality of the event. It any other high level match, each opponent has access to hundreds, sometimes thousands, of game scores played by the other party. Deep Blue certainly had all of Kasparov's prior game scores; Kasparov was foolish to accept the match conditions that kept Deep Blue's games secret. Furthermore, IBM did much in its marketing scheme to present the match as a symmetric competition, and this was unnecessarily misleading to the public because there was non-symmetric secrecy involved.

To this argument Hyatt had also said earlier: does the team know what Kasparov had recently changed in his preparation? No, but the overall play of him is known. Also his particular belief into the magic of mysteries and superstition. The finding of my analyses of the video from 2005 was that someone like Campbell is even now completely unable to realise what the point is for a human chessplayer. Because he argued that Kasparov with his play record was uninteresting because they had created a machine that played just chess and not chess against Kasparov. Now let's assume this isnt a lie, then we come closer to the evaluation of such a match setting. Has a player any chances to understand the chess of a machine in just three games with every color? If he hasnt seen anything else of the newly made robot? Unlikely. So what is the point?
The whole event degenerates to a psycho show because the human in his desperate tries to decrypt a personality walks right into superstition when something happens like the personality switch between round 1 and 2. If he then asks for details and gets impolite responses the whole mindframe is shaken because suddenly people like him, from mankind, are against him, when formerly it was always a sort of collegial togetherness bnetween the team and him as the genial chessplayer. For such a psycho war Kasparov is the worst example among the top players. Because of his weakness, superstition. Especially for a guy like him with his preparations, here without a clue, or worse, without correct clues, the main strength of him was neutralized.

So, why bother at all? What is the point here?

The point is still that such a match setting should have possibly peoven what a machine in chess could do against the best human player and his chess. His chess but not his superstition or psychological weaknesses. So here the organizers allowed to create psychological obstacles in favor of the machine who allegedly just wanted to play chess, not even chess against Kasparov. Ridiculous. It was a psycho war to get to the Prize money. The famous playing dirty of the American sport. That is not British fairness what the Brits had invented as a basis for any sport.

3) Kasparov's baseless innuendo and outright charges of cheating were behavior unacceptable by any standard. In a way, this was even worse than some of Fischer's tirades about Soviet cheating as at least he was correct a few times.

4) The log files produced by the Deep Blue team were poorly formatted and could have been far more informative. I was surprised at how sloppy some of them were, like the entire logging process was a design afterthought. Well formatted and informative logs made available in real time to observers would have properly eliminated much of the controversy. Also, such logs would have been a good contribution to the historical record for other computer chess researchers. Again, we see the unfounded desire for secrecy on the part of the Deep Blue team.

5) The event did more to kill funding and support for computer chess research than any other. The prize money disbursed could have paid for years of projects with chances of good results that would have helped everyone in the field. But instead we have little more than a bunch of old circuit boards in the Smithsonian.

If necessary more about the other points, later.

[Rolf]

My observations on Kasparov vs Deep Blue rematch:

1) Kasparov is an adult, is well experienced, and had already played in one match with Deep Blue. And he had access to advisers, although it's not known if he took any advice. Anyway, he knew the conditions of the rematch and had little right to complain as the Deep Blue team followed the rules.

Rules? Adult behavior? You must have missed my fundamental critic. In such a show the organiser is the one who must supervise the science and good manners aspect of universal fairness, in short, the ethical rules of a civilised society. If say in the 100 meter hurdles final one party were allowed to construct deeper watered holes after every two hurdles and the opposing runner as a favorite for the win accidentaly falls into such a hole you wont accuse him of not having ordered in the contracts that such holes would be forbidden. Hyatt doesnt become tired to claim that in fact a runner must do that too. Ridiculous and odd.

That's ridiculous. If you sign a contract that gives your opponent the right to change the course, then if they do so, it is not _their_ fault, it is _yours_. The match conditions for 1997 were exactly the same as they were for 1996. Kasparov signed on the dotted line for both. It is somewhat like signing a mortgage to buy a house, and then a year later you finally figure out you just can't pay that much per month and are going to default on the loan and lose the house, and then you blame the bank for letting you buy a house you could not afford. You have to accept responsibility for your own actions. Kasparov didn't.

Please Bob, try to be a reasonable scientist for a moment and not just a businessman who has this dollar note blinking in his eyes. Suddenly what seemed to be ridiculous becomes reasonable again. It's all on videotape, no need to change the topic. Campbell screw it.

[sje]

Kasparov's _main_ problem was listening to Friedel, who had him convinced that sparring with the current Fritz would be very similar to playing against DB2.

Simply incredible.

A program challenger should have fought its way to the top to earn the right to a match, just like a human challenger. And that would have meant dozens of games at the least. But IBM, with Kasparov's consent, used its fat wallet to short circuit the process.

Of course, one couldn't have relied upon FIDE to have helped with any of this. But IBM could have financed a number of serious public matches with titled players playing Deep Blue in a process paralleling the regular world championship qualification sequence.

[Dann Corbit]

Kasparov's _main_ problem was listening to Friedel, who had him convinced that sparring with the current Fritz would be very similar to playing against DB2.

Simply incredible.

A program challenger should have fought its way to the top to earn the right to a match, just like a human challenger. And that would have meant dozens of games at the least. But IBM, with Kasparov's consent, used its fat wallet to short circuit the process.

Of course, one couldn't have relied upon FIDE to have helped with any of this. But IBM could have financed a number of serious public matches with titled players playing Deep Blue in a process paralleling the regular world championship qualification sequence.

Do you imagine that Kasparov did not negotiate for the things he thought were important? The same will be true for the IBM team.
Now, I agree to some degree that it would have been very valuable to get Deep Blue games. But how much would it have helped, since IBM was allowed to reprogram Deep Blue between games?

I think Kasparov got bad advice for the match, but he got to choose any advisors he liked so that cannot be blamed on IBM in any way that I can think of.

We could always revise history and make the rules for the match different and possibly change the outcome. But it is too late for that. Wouldn't it be fun to see Deep Blue verses Hydra? How about Hydra, Deep Blue and Rybka 3 on 16 CPUs? Let's toss in the top 3 Fide players for an even more exciting match. Yes, but we won't ever see it. Deep Blue is dismantled, Hydra has gone into a corner and GMs are hard to convince to play against Silicon today.

[bob]

Kasparov's _main_ problem was listening to Friedel, who had him convinced that sparring with the current Fritz would be very similar to playing against DB2.

Simply incredible.

A program challenger should have fought its way to the top to earn the right to a match, just like a human challenger. And that would have meant dozens of games at the least. But IBM, with Kasparov's consent, used its fat wallet to short circuit the process.

Of course, one couldn't have relied upon FIDE to have helped with any of this. But IBM could have financed a number of serious public matches with titled players playing Deep Blue in a process paralleling the regular world championship qualification sequence.

Do you imagine that Kasparov did not negotiate for the things he thought were important? The same will be true for the IBM team.
Now, I agree to some degree that it would have been very valuable to get Deep Blue games. But how much would it have helped, since IBM was allowed to reprogram Deep Blue between games?

I think Kasparov got bad advice for the match, but he got to choose any advisors he liked so that cannot be blamed on IBM in any way that I can think of.

We could always revise history and make the rules for the match different and possibly change the outcome. But it is too late for that. Wouldn't it be fun to see Deep Blue verses Hydra? How about Hydra, Deep Blue and Rybka 3 on 16 CPUs? Let's toss in the top 3 Fide players for an even more exciting match. Yes, but we won't ever see it. Deep Blue is dismantled, Hydra has gone into a corner and GMs are hard to convince to play against Silicon today.

They aren't that hard to convince. Just do as IBM did in 1997 and dangle $1,000,000.00 under his nose. If they had shouted "frog" he would have "jumped"... Problem is, of course, not everyone has $1M laying around where they can offer it to a GM to play a match.