Kasparov vs Deep Blue,what happened?

[MattieShoes]

Of course IBM is concerned with the bottom line. They're not an NPO. Making money is what they DO. The science they support is simply long term ways to make more money. Criticizing them for it seems... odd. Was somebody under the impression that they weren't?

As for not granting a rematch, why would they? They got exactly what they wanted already. They beat the champ. There was nowhere to go but down. The Superbowl champs don't grant rematches either. And this was similar -- a lot of hype and glitz with a game thrown in that is usually worse than many games from the "regular season".

Take another example. Fischer beat the champ, then retired without defending his title. If he had played and lost to Karpov rather than quitting, would people still be putting him on a pedestal? Well, probably they would simply because he was American, but a lot less so.

[Rolf]

Hsu et al. may have been scientists, but this match had nothing to do with science.
When you do science, you leave a legacy, something to improve on, open knowledge to share etc. This left 6 games and lots of comments and gossips. If science would have been the goal, the process would have been more open and the project would not have been dismantled.

Miguel

Sorry, but not true. They published several papers, including a tell-all about singular extensions and other ideas they used... Hsu published papers about his "belle-on-a-chip" design, about the re-design for DB1, for the re-re-design for DB2. The parallel search of DB. They were quite open. IBM was a different story of course...

Employees can hardly overrule their upper management...

Exactly my point. This *match* had nothing to do with science. You brougth up Hsu, I did not. We were talking all the time about IBM.

Miguel

Already in 1997 I informed Bob why the scientists, his friends, spoiled it. Also under dictate of IBM, they were not entitled in the least, to sell out the basics of any science. That you dont mistreat your client and above all the Law that you cant prove something well defined if you then apply something different as the main movens. In other words, to their science it belonged the knowledge that you dont beat the chess of a human player if you psych him out. That is so trivial that it cant be ignored unless people have dishonest motives. For IBM it's correct to say that money doesnt stink.

[Terry McCracken]

The whole thing was ill conceived from the beginning. It was a humongous publicity stunt and GK fell for it, since he thought was a demi-god (he never thought he could lose to force a rematch in the contract, am I right?).

There is very little positive aspects from all this. GK had an infantile behavior and IBM was just despicable. No third match, no log files... the whole thing was theater. Lots of money invested in what? 6 + 6 pgn files and some junk display in the smithsonian. The games were not even good! and some were a terrible spanking or just suicidal. Both, chess and computer chess lost with this match. IBM made a lot of money in the stock market though. That was the bottom line.
Miguel

Very True! Well said!

I also completely agree.
The big question for me is how kasparov was psyched out. Maybe because he could not intimidate the computer.

As for ibm being honest. They said they were interested in the science. If they were so interested in the science why did they dismantle db? Why didn’t db ever play anyone else? In science you need repeatability to be verified as true. This makes me suspicious of the whole match. Why were they so quick to dismantle db? What was really going on behind the guarded doors? I don’t think we will ever know.

This is not the first time ibm tried to pull a stunt like this. In 1962 ibm claimed its checkers program beat a top ranked master checker player. It turned out the human player was a weaker major player. The reason the computer won was because the human blundered. In fact both players had many chances to win because both players blundered many times. Its all documented on page 94 in Schaeffer’s book 1 jump ahead.
Mark

You're right, the science wasn't important to IBM, the rise in stocks was all that mattered to them.

Terry

I disagree. To Hsu, Campbell, Hoane, et. al, the goal was to beat Kasparov. It was _everybody's_ goal from the 70's onward. It was the ultimate goal and was even solidified with the Fredkin prize for the first to beat the current WC in a match. Everyone pursued this. So the "developers" had a goal that was 100% in line with science. They published papers, they participated at nearly every computer chess event that was held, they discussed what they were doing as openly as anyone. The events starting in 1996 were the result of higher-ups recognizing the value of such a victory. And that;'s where things went wrong. But not with the original DT/DB group...

Bob, you don't disagree at all. I was talking about IBM not Hsu, Cambell, Hoane, et. al.

When IBM saw Kasparov was beaten, regardless of the reasons they deemed their job finished and the work of Hsu et. al was over the team got their _Pink Slips_ from the project and IBM saw to it that Deep Blue was dismantled.

Hsu, Cambell, et. al were _not_ IBM.

[Dann Corbit]

Carbon won the 1st
Silicon the 2nd

Must have been a pretty good match, because we are still talking about it over a decade later.

[Dann Corbit]

Well I think IBM acted with honestly and integrity throughout, and any accusation of cheating is totally untrue. Kasparov simply threw a tantrum when he got outplayed by a machine. Its as simple as that for me.
It was in fact this match which got me really interested in computer chess

The whole thing was ill conceived from the beginning. It was a humongous publicity stunt and GK fell for it, since he thought was a demi-god (he never thought he could lose to force a rematch in the contract, am I right?).

There is very little positive aspects from all this. GK had an infantile behavior and IBM was just despicable. No third match, no log files... the whole thing was theater. Lots of money invested in what? 6 + 6 pgn files and some junk display in the smithsonian. The games were not even good! and some were a terrible spanking or just suicidal. Both, chess and computer chess lost with this match. IBM made a lot of money in the stock market though. That was the bottom line.

Miguel

All is reasonably correct except for the log files. They were released and I still have them somewhere if you have not seen them. I got them from IBM's web site years ago. The log snippet for a couple of the moves in question were released and discussed at length in r.g.c.c at the time of the match...

The log files should have been made public the next day after the match ended. In fact, any true scientist would have shown them to GK immediately after he complained. But maybe it was a business, sport, or legal decision. A scientist is not interested in beating a world champion that is not playing at its best.

Miguel

This was a science experiment only for those who generated the machine and programs. The owner of the process (IBM) was doing it striclty for P.R.

Hence, any decisions about strategic actions to take would have to run their gauntlet through the boardroom and get all the bigwhigs to nod in approval.

That meant that nothing unexpected would happen quickly.

[bob]

The whole thing was ill conceived from the beginning. It was a humongous publicity stunt and GK fell for it, since he thought was a demi-god (he never thought he could lose to force a rematch in the contract, am I right?).

There is very little positive aspects from all this. GK had an infantile behavior and IBM was just despicable. No third match, no log files... the whole thing was theater. Lots of money invested in what? 6 + 6 pgn files and some junk display in the smithsonian. The games were not even good! and some were a terrible spanking or just suicidal. Both, chess and computer chess lost with this match. IBM made a lot of money in the stock market though. That was the bottom line.
Miguel

Very True! Well said!

I also completely agree.
The big question for me is how kasparov was psyched out. Maybe because he could not intimidate the computer.

As for ibm being honest. They said they were interested in the science. If they were so interested in the science why did they dismantle db? Why didn’t db ever play anyone else? In science you need repeatability to be verified as true. This makes me suspicious of the whole match. Why were they so quick to dismantle db? What was really going on behind the guarded doors? I don’t think we will ever know.

This is not the first time ibm tried to pull a stunt like this. In 1962 ibm claimed its checkers program beat a top ranked master checker player. It turned out the human player was a weaker major player. The reason the computer won was because the human blundered. In fact both players had many chances to win because both players blundered many times. Its all documented on page 94 in Schaeffer’s book 1 jump ahead.
Mark

You're right, the science wasn't important to IBM, the rise in stocks was all that mattered to them.

Terry

I disagree. To Hsu, Campbell, Hoane, et. al, the goal was to beat Kasparov. It was _everybody's_ goal from the 70's onward. It was the ultimate goal and was even solidified with the Fredkin prize for the first to beat the current WC in a match. Everyone pursued this. So the "developers" had a goal that was 100% in line with science. They published papers, they participated at nearly every computer chess event that was held, they discussed what they were doing as openly as anyone. The events starting in 1996 were the result of higher-ups recognizing the value of such a victory. And that;'s where things went wrong. But not with the original DT/DB group...

Hsu et al. may have been scientists, but this match had nothing to do with science.
When you do science, you leave a legacy, something to improve on, open knowledge to share etc. This left 6 games and lots of comments and gossips. If science would have been the goal, the process would have been more open and the project would not have been dismantled.

Miguel

Sorry, but not true. They published several papers, including a tell-all about singular extensions and other ideas they used... Hsu published papers about his "belle-on-a-chip" design, about the re-design for DB1, for the re-re-design for DB2. The parallel search of DB. They were quite open. IBM was a different story of course...

Employees can hardly overrule their upper management...

Exactly my point. This *match* had nothing to do with science. You brougth up Hsu, I did not. We were talking all the time about IBM.

Miguel

You said "when you do science" and they certainly did their fair share of that. No match or tournement ever played was about "science". CCTs. WCCCs. There's absolutely no science in those events either, so I am not sure I see your point...

The DB team certainly published whatever they did...

[bob]

Hsu et al. may have been scientists, but this match had nothing to do with science.
When you do science, you leave a legacy, something to improve on, open knowledge to share etc. This left 6 games and lots of comments and gossips. If science would have been the goal, the process would have been more open and the project would not have been dismantled.

Miguel

Sorry, but not true. They published several papers, including a tell-all about singular extensions and other ideas they used... Hsu published papers about his "belle-on-a-chip" design, about the re-design for DB1, for the re-re-design for DB2. The parallel search of DB. They were quite open. IBM was a different story of course...

Employees can hardly overrule their upper management...

Exactly my point. This *match* had nothing to do with science. You brougth up Hsu, I did not. We were talking all the time about IBM.

Miguel

Already in 1997 I informed Bob why the scientists, his friends, spoiled it. Also under dictate of IBM, they were not entitled in the least, to sell out the basics of any science. That you dont mistreat your client and above all the Law that you cant prove something well defined if you then apply something different as the main movens. In other words, to their science it belonged the knowledge that you dont beat the chess of a human player if you psych him out. That is so trivial that it cant be ignored unless people have dishonest motives. For IBM it's correct to say that money doesnt stink.

my "friends" didn't spoil anything. They developed an amazing, for the time, chess playing box. They beat Grandmasters right and left. They won the second and third stages of the Fredkin prize. The final stage by beating Kasparov. I can think of several reasons to discontinue the project after the second match.

1. they won

2. they had to put up with a ton of Crap from Kasparov and friends, when they should have been sitting back and enjoying the fruits of a job well done...

3. IBM didn't psych _anybody_ out. Kasparov was responsible for that and did a damn fine job of it in fact...

[bob]

Well I think IBM acted with honestly and integrity throughout, and any accusation of cheating is totally untrue. Kasparov simply threw a tantrum when he got outplayed by a machine. Its as simple as that for me.
It was in fact this match which got me really interested in computer chess

The whole thing was ill conceived from the beginning. It was a humongous publicity stunt and GK fell for it, since he thought was a demi-god (he never thought he could lose to force a rematch in the contract, am I right?).

There is very little positive aspects from all this. GK had an infantile behavior and IBM was just despicable. No third match, no log files... the whole thing was theater. Lots of money invested in what? 6 + 6 pgn files and some junk display in the smithsonian. The games were not even good! and some were a terrible spanking or just suicidal. Both, chess and computer chess lost with this match. IBM made a lot of money in the stock market though. That was the bottom line.

Miguel

All is reasonably correct except for the log files. They were released and I still have them somewhere if you have not seen them. I got them from IBM's web site years ago. The log snippet for a couple of the moves in question were released and discussed at length in r.g.c.c at the time of the match...

The log files should have been made public the next day after the match ended. In fact, any true scientist would have shown them to GK immediately after he complained. But maybe it was a business, sport, or legal decision. A scientist is not interested in beating a world champion that is not playing at its best.

Miguel

I fail to see how paranoia on the part of Kasparov constitutes wrong-doing on the part of IBM. He didn't ask for the log files in a polite manner. And had that been me, I would have told him exactly where he could go, and if I chose to give him the logs, exactly where I would shove them.

[Rolf]

This was a science experiment only for those who generated the machine and programs. The owner of the process (IBM) was doing it striclty for P.R.

Hence, any decisions about strategic actions to take would have to run their gauntlet through the boardroom and get all the bigwhigs to nod in approval.

That meant that nothing unexpected would happen quickly.

This all sounds very reasonable but is it also sound? I had the same exchange with Bob in 1997. His position always was that scientists ALWAYS are depending of the money dictates. Which is apparently true also in my eyes. However here the question was, what machine can beat the World Champ for the first time in a (very short) match. Called Fredkin Prize what could be won.

Now my counter position was always that of course you can win in chess ALSO with off-board influences. My strict position would that always prove that the machine won against the human player - IF the operaters and advisors and contact persons for Kasparov treated him with a certain disrespect? My opinion was that NO, in such cases the result would have been regarded as irregular and invalid. So in truth no win of the Fredking Prize. And I argued that alone for that possibility the scientists should have shown interest that Kasparov should be treated with respect and this way the team&IBM would win back the initiative.

The event would be connected for all times with the deceived Kasparov, so that in the long run it would speak against the IBM side plus Hsu et al.

After Fischer is dead I wish to make another point. In 1972 we had a similar situation like 25 yeas later. Fischer showed disrespect for the Soviet side. And normally Spassky should have won the game by Fischer's forfeit. Even I personally went almost mad when Bobby refused to play the third game. I wrote a telegram from Switzerland that I would lose all my hair if he didnt play because I had running a bet against some Americans and Canadians in a International Youth Hostel there.

Today after Fischer is dead now, I must reconsider my position. What he did was normally impossible to tolerate for Spassky and he should have refused to further play Fischer. But Spassky believed wrongly in his superiority and was basically psyched out by Fischer.

Because if you ever played serious chess then you know that you can become superstitious. And if the opponent always gets his extra rights then this can suck you out. Therefore Spassky should have refused to continue to play.

And for the same reasons Kasparov in 1997 should have stopped playing if he couldnt fully trust the IBM team. Because Bob showed more than enough possibilities to cheat in such a constellation. So either you play your chess and trust the engineers or you know you are superstituous and then you shouldnt go into such a match. - So yes, in a way I begin to revise my position. Since I dont know any chess proof for Kasparov being cheated I vote for suspension of the decision. Maybe it will come out in future maybe not. Until now I fully believed in Kasparov but he couldnt make a watertight case. Again, perhaps it will come out but in the meantime it makes no sense to speculate.

Another question is of course if the scientists shouldnt have supported Kasparov when he came into the often described crisis of trust. Because what then happened wasnt a real match if Kasparov had almost no chance to recover in such a short run. He just was psyched out. The result is meaningless.

Unfortunately the whole event can never more be repeated because now the machines are far too strong in chess although they still cannot compete IMO in classical chess because of - as I say - unfair conditions with the tools, and - as Bob says - because of the fear to lose of the human players.

I am fully satisfied with these fantastic training tools and training partners in chess.

Footnote:

I would have wanted that many of the points would have been found in open debates here because what are we here discussing for? For all we shouldnt defamate each other.

Friendly regards to all readers and members.

[bob]

This was a science experiment only for those who generated the machine and programs. The owner of the process (IBM) was doing it striclty for P.R.

Hence, any decisions about strategic actions to take would have to run their gauntlet through the boardroom and get all the bigwhigs to nod in approval.

That meant that nothing unexpected would happen quickly.

This all sounds very reasonable but is it also sound? I had the same exchange with Bob in 1997. His position always was that scientists ALWAYS are depending of the money dictates. Which is apparently true also in my eyes. However here the question was, what machine can beat the World Champ for the first time in a (very short) match. Called Fredkin Prize what could be won.

Now my counter position was always that of course you can win in chess ALSO with off-board influences. My strict position would that always prove that the machine won against the human player - IF the operaters and advisors and contact persons for Kasparov treated him with a certain disrespect? My opinion was that NO, in such cases the result would have been regarded as irregular and invalid. So in truth no win of the Fredking Prize. And I argued that alone for that possibility the scientists should have shown interest that Kasparov should be treated with respect and this way the team&IBM would win back the initiative.

The event would be connected for all times with the deceived Kasparov, so that in the long run it would speak against the IBM side plus Hsu et al.

After Fischer is dead I wish to make another point. In 1972 we had a similar situation like 25 yeas later. Fischer showed disrespect for the Soviet side. And normally Spassky should have won the game by Fischer's forfeit. Even I personally went almost mad when Bobby refused to play the third game. I wrote a telegram from Switzerland that I would lose all my hair if he didnt play because I had running a bet against some Americans and Canadians in a International Youth Hostel there.

Today after Fischer is dead now, I must reconsider my position. What he did was normally impossible to tolerate for Spassky and he should have refused to further play Fischer. But Spassky believed wrongly in his superiority and was basically psyched out by Fischer.

Because if you ever played serious chess then you know that you can become superstitious. And if the opponent always gets his extra rights then this can suck you out. Therefore Spassky should have refused to continue to play.

And for the same reasons Kasparov in 1997 should have stopped playing if he couldnt fully trust the IBM team. Because Bob showed more than enough possibilities to cheat in such a constellation. So either you play your chess and trust the engineers or you know you are superstituous and then you shouldnt go into such a match. - So yes, in a way I begin to revise my position. Since I dont know any chess proof for Kasparov being cheated I vote for suspension of the decision. Maybe it will come out in future maybe not. Until now I fully believed in Kasparov but he couldnt make a watertight case. Again, perhaps it will come out but in the meantime it makes no sense to speculate.

Another question is of course if the scientists shouldnt have supported Kasparov when he came into the often described crisis of trust. Because what then happened wasnt a real match if Kasparov had almost no chance to recover in such a short run. He just was psyched out. The result is meaningless.

Unfortunately the whole event can never more be repeated because now the machines are far too strong in chess although they still cannot compete IMO in classical chess because of - as I say - unfair conditions with the tools, and - as Bob says - because of the fear to lose of the human players.

I am fully satisfied with these fantastic training tools and training partners in chess.

Footnote:

I would have wanted that many of the points would have been found in open debates here because what are we here discussing for? For all we shouldnt defamate each other.

Friendly regards to all readers and members.

Rolf,

Kasparov was treated to a ton of respect. I believe he won $400,000 in the _first_ match, and he could have won $700,000 in the second match had he won, but he again took home $400,000 for that match. A cool $800,000 dollars of "respect". Kasparov negotiated the contract in great detail, the negotiations dragged on and on. And then, once the match starts, he decides that the terms are not so agreeable and he begins to complain, insult, etc.

If he got psyched out, that was _his_ fault. He has no one to blame but himself. He resigned a drawn game, all by himself. He played a poor opening in a critical round. All by himself. He really has no one to blame, but himself... Trying to blame IBM is just excuse-making...