- Blind mode tutorial
lichess.org
Donate

IBM Summary on Deep Blue

#1

"On May 11, 1997, an IBM computer called IBM ® Deep Blue ® beat the world chess champion after a six-game match: two wins for IBM, one for the champion and three draws. The match lasted several days and received massive media coverage around the world. It was the classic plot line of man vs. machine. Behind the contest, however, was important computer science, pushing forward the ability of computers to handle the kinds of complex calculations needed to help discover new medical drugs; do the broad financial modeling needed to identify trends and do risk analysis; handle large database searches; and perform massive calculations needed in many fields of science."

"On May 11, 1997, an IBM computer called IBM ® Deep Blue ® beat the world chess champion after a six-game match: two wins for IBM, one for the champion and three draws. The match lasted several days and received massive media coverage around the world. It was the classic plot line of man vs. machine. Behind the contest, however, was important computer science, pushing forward the ability of computers to handle the kinds of complex calculations needed to help discover new medical drugs; do the broad financial modeling needed to identify trends and do risk analysis; handle large database searches; and perform massive calculations needed in many fields of science."
#2

May 11, 1997 was indeed the fall of humanities dominance over chess. With the rapid development of technology, not only could robots take up low-skilled jobs, but also chess, where people thought creativity was the key to success, and that only humans have creativity and so computers will never succeed. However, now with computers like stockfish and komodo 9 defeating strong players like Nakamura even when handicapped a pawn, is testimony to the strength of technology

May 11, 1997 was indeed the fall of humanities dominance over chess. With the rapid development of technology, not only could robots take up low-skilled jobs, but also chess, where people thought creativity was the key to success, and that only humans have creativity and so computers will never succeed. However, now with computers like stockfish and komodo 9 defeating strong players like Nakamura even when handicapped a pawn, is testimony to the strength of technology
#3

" Everything has a price" (Thomas Mann)

The root of all evil was Cray Blitz (1968?). Not to forget it's successor for the masses: Crafty. Some say this is still the most importent chess software written ever.

See also https://chessprogramming.wikispaces.com/

" Everything has a price" (Thomas Mann) The root of all evil was Cray Blitz (1968?). Not to forget it's successor for the masses: Crafty. Some say this is still the most importent chess software written ever. See also https://chessprogramming.wikispaces.com/
#4

I doubt Deep Blue beat Kasparov at that point in time without human aid.

  1. The printouts of what the computer was thinking were never shown publicly or to Kasparov

  2. Deep Blue was dismantled immediately following the match.

  3. No engines were anywhere near that strong at that point in time.

  4. A rematch was declined

  5. IBM stock rose significantly the day after it beat Kasparov

  6. A human being cutting bad lines from an engine can make the engine play much stronger.

I doubt Deep Blue beat Kasparov at that point in time without human aid. 1) The printouts of what the computer was thinking were never shown publicly or to Kasparov 2) Deep Blue was dismantled immediately following the match. 3) No engines were anywhere near that strong at that point in time. 4) A rematch was declined 5) IBM stock rose significantly the day after it beat Kasparov 6) A human being cutting bad lines from an engine can make the engine play much stronger.
#5

@lurarose

  1. what should have been shown? what you see from an engine today? i doubt the software returned any sort of helpful information because it was built for a very specific purpose that not include user friendly output
  2. if they had faked the result, why would they need to dismantle it? it's not like they were loaning it out at the library for people to hook up and try out
  3. a lot of research was put into deep blue's construction, and deep blue was the first one to implement these advances in engine theory.
  4. why was IBM obligated to a rematch?
  5. of course it did, it showed off ibm's engineering prowess. would it have not increased if Deep Blue wasn't "fake"?
  6. can it? an engine already prunes lines (usually), i don't see why a human also has to

also the obvious counter (for me):

  1. who could have possibly helped? they would obviously need to be a strong GM to engineer moves that would have beat Kasparov, and chess players generally don't act so slimy at that level.
@lurarose 1) what should have been shown? what you see from an engine today? i doubt the software returned any sort of helpful information because it was built for a very specific purpose that not include user friendly output 2) if they had faked the result, why would they need to dismantle it? it's not like they were loaning it out at the library for people to hook up and try out 3) a lot of research was put into deep blue's construction, and deep blue was the first one to implement these advances in engine theory. 4) why was IBM obligated to a rematch? 5) of course it did, it showed off ibm's engineering prowess. would it have not increased if Deep Blue wasn't "fake"? 6) can it? an engine already prunes lines (usually), i don't see why a human also has to also the obvious counter (for me): 7) who could have possibly helped? they would obviously need to be a strong GM to engineer moves that would have beat Kasparov, and chess players generally don't act so slimy at that level.
#7

@ #4 - "No engines were anywhere near that strong at that point in time." - what engines? So far as I know, Deep Blue was the first machine to be built solely to be able to beat a GM at chess... all other "chess computers" at the time were for the average person, therefore they weren't that powerful.

@ #4 - "No engines were anywhere near that strong at that point in time." - what engines? So far as I know, Deep Blue was the first machine to be built solely to be able to beat a GM at chess... all other "chess computers" at the time were for the average person, therefore they weren't that powerful.
#8

@ianremsen

  1. None of the thought process of deep blue was recorded or shown. The operators could see what deep blue thought was best, and how it changed it's mind, "What is the strongest line at this point in time" according to the machine. Nothing like that was shown, just the end result. It's the equivalent of a chef showing you the food they have made, but not how they cooked it up.

  2. Dismantling it is a way of getting rid of the evidence. It's the equivalent of putting a man on mars... then packing the bags, going home, and moving on never to venture into space again. IBM has done NOTHING with chess since this event.

  3. This has not been proven by IBM.

  4. Kasparov defeated deep blue the year before. He immediately offered a rematch which they accepted. This time they totally declined him any chance of redemption.

  5. Ya, they were able to put together a team to beat Kasparov at chess.

  6. Human's can prune lines that an engine can not see as bad. If the computer is going on some sort of pawn grubbing strategy, and leaving the king in danger... Even I a player of around 2000 rating can clearly see this. If I were put behind deep blue and saw it doing that... guess what I cut that line. Push a button and it stops doing that, it's next most powerful line comes to the top which is the truly stronger line. They would not need Karpov there to be the line cutter. Just a player of my strength could boost the engine performance drastically, although yes Karpov or another GM would probably trim lines more accurately than I, I would be able to do an adequate job to improve the strength of the engine significantly.

  7. Like I just said... You don't need a powerful grandmaster. A 2000-2200 player would suffice. Even a properly instructed 1800 player with the instructions of "It's going after pawns, and leaving the king in danger. Delete that line." would be able to improve the machine a lot. It's simply a matter of getting rid of the bad lines that computers at this point in time of understanding were bad. Also, they did have several GM working along side the deep blue team any of which would be more than capable of doing this accurately.

@ianremsen 1) None of the thought process of deep blue was recorded or shown. The operators could see what deep blue thought was best, and how it changed it's mind, "What is the strongest line at this point in time" according to the machine. Nothing like that was shown, just the end result. It's the equivalent of a chef showing you the food they have made, but not how they cooked it up. 2) Dismantling it is a way of getting rid of the evidence. It's the equivalent of putting a man on mars... then packing the bags, going home, and moving on never to venture into space again. IBM has done NOTHING with chess since this event. 3) This has not been proven by IBM. 4) Kasparov defeated deep blue the year before. He immediately offered a rematch which they accepted. This time they totally declined him any chance of redemption. 5) Ya, they were able to put together a team to beat Kasparov at chess. 6) Human's can prune lines that an engine can not see as bad. If the computer is going on some sort of pawn grubbing strategy, and leaving the king in danger... Even I a player of around 2000 rating can clearly see this. If I were put behind deep blue and saw it doing that... guess what I cut that line. Push a button and it stops doing that, it's next most powerful line comes to the top which is the truly stronger line. They would not need Karpov there to be the line cutter. Just a player of my strength could boost the engine performance drastically, although yes Karpov or another GM would probably trim lines more accurately than I, I would be able to do an adequate job to improve the strength of the engine significantly. 7) Like I just said... You don't need a powerful grandmaster. A 2000-2200 player would suffice. Even a properly instructed 1800 player with the instructions of "It's going after pawns, and leaving the king in danger. Delete that line." would be able to improve the machine a lot. It's simply a matter of getting rid of the bad lines that computers at this point in time of understanding were bad. Also, they did have several GM working along side the deep blue team any of which would be more than capable of doing this accurately.
#9
  1. None of the thought process of deep blue was recorded or shown.

Those are the logs - a computer doesn't "think", it just brute-forces, then evaluates each position, checks which is best, and plays the relevant move. The only thing "deeper" than the logs would be things like code for the operating system, which would only be useful if Deep Blue was rebuilt. OK, there wouldn't be a thing in the log saying "Human intervenes at this point", but what are you looking for in the logs which proves that a human nudged Deep Blue in the right direction?

1) None of the thought process of deep blue was recorded or shown. Those are the logs - a computer doesn't "think", it just brute-forces, then evaluates each position, checks which is best, and plays the relevant move. The only thing "deeper" than the logs would be things like code for the operating system, which would only be useful if Deep Blue was rebuilt. OK, there wouldn't be a thing in the log saying "Human intervenes at this point", but what are you looking for in the logs which proves that a human nudged Deep Blue in the right direction?
#10

There's no real evidence of any wrongdoing or "funny business" by the IBM team; there are only these vague, entirely circumstantial, not-very-convincing reasons based on some assumptions about how engines couldn't possibly play, in turn based on just a couple moves of chess.

The most probable scenario is that this was all Kasparov being, well, Kasparov. He just isn't a gracious loser by any stretch of the imagination and coming up with some excuses to explain away a loss is right up his alley.

Commercial engines running on consumer-grade PCs were already quite strong at that time. It's true they wouldn't have been expected to beat Kasparov in classical chess, but it's not so surprising that on massive dedicated hardware the gap could be closed, at least for a single short match (aside from the ridiculous blunder in game 6, analysis with modern engines has indicated that Deep Blue didn't really play any stronger than Kasparov; if anything, Kasparov's play was generally stronger).

It's not unusual for a slightly stronger player to occasionally lose a short match.

So, we can either just accept that the world champion lost a short match against a massive dedicated machine in 1997, or buy into some very speculative conspiracy theories, but I'm just not sure what the point is of the latter.

Also, it's been pointed out several times that modern engines very quickly find the alleged "cheating/human/impossible-for-engines-to-find" moves, so it's not all unreasonable to think that Deep Blue could have found it on its own on its massive dedicated hardware (again, they were just a couple moves; even the crappiest players/engines occasionally find moves beyond their level).

To me, it just seems much more likely that Deep Blue found (either by its normal means, or as has also been suggested, as a result of a bug) this strong sequence (axb5 and Be4!) instead of going for the immediate win of material, and Kasparov was shocked that a computer didn't blindly go for a win of material.

When that resulted in a loss, he just started trying to explain it away, because he's not a gracious loser and has an enormous ego.

To me, that seems much more likely than the alternative conspiracy theory (and aside from the seeming-to-me, it's important to reiterate that all of this hullabaloo is based on a combination of two moves of chess and some very questionable assumptions about what Deep Blue couldn't possibly do, scant evidence in the best of light), but we're all entitled to our opinions, I suppose.

Cheers!

There's no real evidence of any wrongdoing or "funny business" by the IBM team; there are only these vague, entirely circumstantial, not-very-convincing reasons based on some assumptions about how engines couldn't possibly play, in turn based on just a couple moves of chess. The most probable scenario is that this was all Kasparov being, well, Kasparov. He just isn't a gracious loser by any stretch of the imagination and coming up with some excuses to explain away a loss is right up his alley. Commercial engines running on consumer-grade PCs were already quite strong at that time. It's true they wouldn't have been expected to beat Kasparov in classical chess, but it's not so surprising that on massive dedicated hardware the gap could be closed, at least for a single short match (aside from the ridiculous blunder in game 6, analysis with modern engines has indicated that Deep Blue didn't really play any stronger than Kasparov; if anything, Kasparov's play was generally stronger). It's not unusual for a slightly stronger player to occasionally lose a short match. So, we can either just accept that the world champion lost a short match against a massive dedicated machine in 1997, or buy into some very speculative conspiracy theories, but I'm just not sure what the point is of the latter. Also, it's been pointed out several times that modern engines very quickly find the alleged "cheating/human/impossible-for-engines-to-find" moves, so it's not all unreasonable to think that Deep Blue could have found it on its own on its massive dedicated hardware (again, they were just a couple moves; even the crappiest players/engines occasionally find moves beyond their level). To me, it just seems much more likely that Deep Blue found (either by its normal means, or as has also been suggested, as a result of a bug) this strong sequence (axb5 and Be4!) instead of going for the immediate win of material, and Kasparov was shocked that a computer didn't blindly go for a win of material. When that resulted in a loss, he just started trying to explain it away, because he's not a gracious loser and has an enormous ego. To me, that seems much more likely than the alternative conspiracy theory (and aside from the seeming-to-me, it's important to reiterate that all of this hullabaloo is based on a combination of two moves of chess and some very questionable assumptions about what Deep Blue couldn't possibly do, scant evidence in the best of light), but we're all entitled to our opinions, I suppose. Cheers!

This topic has been archived and can no longer be replied to.