Deep Blue vs Kasparov: How a computer beat best chess player in the world - BBC News

Twenty years ago IBM’s Deep Blue defeated previously unbeaten chess grandmaster Gary Kasparov. Its designers tell the BBC how they won and what it means for computing.
Produced by the BBC’s Franz Strasser.
Please subscribe HERE bit.ly/1rbfUog
World In Pictures • WORLD IN PICTURES
Big Hitters • MUST SEE VIDEOS
Just Good News • Just good news

[@Roman-id4vn]

And yet the CS GO bot on my team continues to rush mid with a glock 18 every round

[@pwncakesgaming]

Meow Dog lmao

[@AhmedKMoustafa2]

LAMO

[@explosivemodesonicmauricet1597]

They have Intelligent AI in DOTA 2 tho.

[@AnkurKumar-ur1cj]

And buy a Negev in an important round.

[@Nemanjaftn]

bro hahhahaha not the kind of comment i would expect here,so funny

[@radikusmanov7574]

I hope they congratulated Garry Kasparov also. Not as a loser, but as key participant of new scientific experiment.

[@radikusmanov7574]

You are wrong. The gunny still competes. Look for plays from this year in KZbin.

[@MrSupernova111]

Nothing Edgar said was incorrect. Kasparov retired years ago and was giving speeches before returning to competitive chess this year. Don't be so quick to jump into conclusions.

[@kherossilverlight8400]

It's almost always underlooked how in the timelapse between the first match (which Kasparov won) and the rematch, Deep Blue was specifically designed to beat Kasparov. Meanwhile, Kasparov was denied access to any of the "new" Deep Blue matches, leaving him at a clear informational disadvantaje.

[@kebabkebob7808]

Cope

[@rakshithanand8262]

@@kebabkebob7808 It's not really copium as such. Kasparov was at a disadvantage in terms of preparation AND the pressure of playing a machine that feels no fear or stress when one of his best strength was the sheer terror he inspired. He was also denied a third set of matches after deep blue won because IBM was not confident in a repeat victory. And he has played against stronger computers since and drawn consistently. Of course, nowadays he is handily outclassed by even mid-range chess apps.

[@M414-q6o]

​@@rakshithanand8262 and how does that change the fact that he lost? Deep blue was still the first computer to beat the world champion. none of your execuses are relevant to this fact

[@Asgoga]

​@@M414-q6oCope more

[@Hury209]

okay then if you think Kasparov has any chance against AlphaZero, then let us see how he does against it.. hehehe :)

[@trixbrix554]

Honest Opinion:- This was an attempt to insult, NOT an attempt to challenge

[@PackerBronco]

The computer did NOT beat Kasparov. After Gary cleaned Deep Blue's clock in Game 1, the programmers went back and made some modifications to the computer program. In other words, human intelligence intervened on Deep Blue's behalf based on what the programmers observed about Gary's strategy. What should have happened is that once the match had begun, Deep Blue was on its own.

[@panama2468]

Wow, that's such a clear violation of ethics

[@omeven5785]

it doesn't work like that, they couldn't adjust deep blue to only work against kasparov, they just polished their code like a good programmer should do. And by the way, a human player would do exactly that, adjust their strategy based on previous experiences with this opponent.

[@PackerBronco]

@@omeven5785 But that's my point. The HUMAN programmers were adjusting the code based on what they observed from the first game. So it was HUMAN intervention to respond to what they saw Gary do in Game 1. It was not Kasparov vs. Deep Blue, it was Kasparov vs. Deep Blue + A Team of Programmers AND chess grandmasters acting as consultants. To be a true test, IBM should not have been allowed to change anything once the match had begun. If that had to tweak something it should have been for a hypothetical 3rd match.

[@jpdixon]

So? A Human programmed Deep Blue so that means that the computer never won because a Human made it.

[@mguddeti]

@@PackerBroncowho do you think made Deep Blue, not programmers?

[@KaPowProduction]

It was a game thrown. Because Kasparov has already won and IBM kept on improving Deep Blue, if Kasparov wins again, they will never stop until IBM finds a way to beat him. So, Kasparov stopped it by losing, It worked.

[@trixbrix554]

BEST REPLY

[@BoleDaPole]

kasporov invested heavily into IBM before the match.

[@prismarinestars7471]

@@BoleDaPole Really? If so that’s genius

[@gatomeong]

@@BoleDaPole holy shit that's a genius move. no wonder why he's one of the best chess players

[@VvC_Lad]

@@prismarinestars7471 he made a company and it was failing I think then IBM took advantage of this and made a contract with Gary to improve their chess engine and gain more money. After the 1st game of gary kasparov with deep blue, ibm stocks increased 3x then they had a rematch where they put gary at a significant disadvantage. They won and earned a lot or money after that. Idk but imo gary was the one exploited by IBM but hey if it wasn’t for that we wouldn’t have good chess engines that helped improve chess.

[@hutterite1]

IBM won that match, not on the board, but off the board. One important condition to policy was they were allowed to tweak, fix and adjust the machine between games. Even during the game. Kasparov was being thrown something different every game. He lost due to pressure; but it had been in the fine print

[@panama2468]

That's a disingenuous way of phrasing what happened. Just bc the cheated legally doesn't mean it was OK.

[@freshlymemed5680]

​​@@panama2468how is it cheating? In the same way a human can adapt from their experiences in the previous match the machine was allowed the same. It just wasnt capable of adapting on the fly by itself during that time while we have chess AIs on our phones that can do just that now. Its definitely not the most graceful way of doing things but its still undeniably impressive.

[@MultiThunder93]

He was expecting logical moves IBM in the end: (failsafe activated) random moves not logical. Kasparov: pikachu

[@jrak4071]

Gary was so good he had an affect on people which made them afraid to beat him, incredible.

[@Krrrimmi]

I only read this on newspaper Sports section back in Feb. '96. It was huge talk among my chess player classmates back in highdchool. Thanx for uploading this. 👑

[@PunksterOS]

I remember this... I'm still young, damn it!

[@sontungle2641]

But after the Deep Blue beat Gary Kasparov, Kasparov find out that IBM was cheating so Kasparov want to rematch but IBM regret and Deep Blue project was cancelled.

[@cattleherder1912]

How exctly would they cheat in a chess game?

[@nothingtoseehere8865]

Son Tung Le there isn’t any evidence to prove cheating, he was just salty

[@sapien153]

What is cheating anyway ?? The computer can do whatever it wants by rules of the game. How can it cheat then ? By getting assistance from a human ? 😂

[@sapien153]

@@jeudiballsl5518 Why is changing the program cheating ? It's well within the rules..

[@Cnut_the_grape]

How did they cheat? Computer assistance? Ultron hiding under the table?

[@Pdiddy2400]

Does not matter how good you become, your fellow man will always find a way to push you into the dirt.

[@karis7539]

Me here after I wrote minimax algorithm for tic tac toe

[@marccasals6366]

HAHAHHAHAHAHA I'M EXACTLY HERE FOR THE SAME REASON

[@LordlyRecords]

Same but for the Gomoku game

[@MonkeyDIvan]

1:53 Is that Jim Lee's second coming?

[@anwitmondal6417]

One russian guy vs a multi national company.

[@alkan3191]

its ironic how Dr.Stone always lead me to searching stuff like this

[@kato_dsrdr]

The amazing thing is that no matter which of them win, it still means that humans are damn awesome... Imagine creating a machine that can play equally against the best human or having a human that can play against the best machine.. It's both awesome.

[@philosophiaentis5612]

Today no human can play against the best machine.

[@sufficientmagister9061]

​@@philosophiaentis5612 Wait until AI systems become conscious; humans will become obsolete in many fields of study & creativity.

[@DoubleVanimation]

People accused deep blue of using a human Nowdays we accuse people of using a chess engine

[@willypro4949]

I wanna see IBM Deep Blue vs Google Deep Mind. That will truly be the match of the millennium

[@Rhannmah]

huh? It would be an absolute beatdown, like a grown man fighting a child. AlphaZero already beats Stockfish 100-0, which is the CURRENT best chess computer player in the world. It is MANY times more powerful and optimized than Deep Blue ever was, and it gets DESTROYED by Deepmind's AlphaZero.

[@willypro4949]

@@Rhannmah but it would be interesting, not for the sake of seeing an old computer get beat up by an AI, just to see if Deep Blue can fight to the end or at least see the conditions that will put Deep Blue in to surrending the match

[@Nemsesis3624]

@@Rhannmah Dude, wtf are you talking about? In which world Alpha Zero beats stockfish 100:0? Alpha Zero only beat Stockfish 8 which was limited by move time and not allowed using it's opening book. Stockfish 11 is rated several 100 Elo points above alpha zero and is TCEC Champion two times in a row now and therefore the best chess engine in the world.

[@bijeshamatya8090]

human phycology comes to play at high stakes gameplay so it might be possible that he was completely broken down. Remember Chess is not just gameplay of strategy human factors also comes to play.So, hurrah to deep blue

[@slinkerdeer]

​@@idiotsopinions Thing about stupid people is they aren't intelligent enough to realise the limits of their own knowledge. So, they lack the intelligence to know they don't have the knowledge. Pretty much a deadly circle that can only be broken by facts and evidence from someone who actually knows what they are gabbing on about

[@nawazeeshali4340]

This match was truly historic.

[@cringepatrol6393]

This sounds like a movie “I was the best… until the computers beat me”

[@Lemonsieur-m4m]

It is important to realize that Kasparov didn't play against a machine but against 5 or 6 masters (and one over powerful machine) who where focused on one board. You cannot fight 6 men at once. So it is exceptional that he won the first round against the world to put it mildly.

[@sharathvenkatesh]

How many people here know that the computer crashed several times during the matches?

[@AbeOrtiz88]

Lmao it pulled a Bobby fisher move.

[@dreamlifter7569]

Really?

[@PackerBronco]

If so, I hope it was on the clock during the reboot.

[@toyclub721]

@@AbeOrtiz88 This was planned to affect Kasparov's psychology.

[@MiguinGineer]

Who went here after reading Dr. Stone's chapter 213

[@Jesusandbible]

The fact Deep Blue "missed" a mating combination proved it was a mega computer PLUS humans versus Kasparov. Can you imagine a match between two grandmasters each of which had a different super computer to help him, or Magnus Carlsen and AlphaZero versus....... someone? So they NEVER WON THAT MATCH. They cheated.

[@thegloriousryius]

Sounds like they programmed the computer to beat him.

[@vikhyath.]

There needs to Deep Blue vs. The latest version of AlphaZero. Want to watch AlphaZero eat Deep Blue alive.

[@gurugeorge]

How is it not just a bunch of clever professors playing beating a Chess master using a computer as a tool?

[@warsinaction2450]

gurugeorge Because of the complexity of a postion in chess, it is very hard to calculate well. It was an incredible achievement to do it. What got him is computers play very weird. They might sac a queen to come out of a 10 move sequence a pawn ahead

[@alexandersapir9079]

Actually, Calculation is the easy part. The hard part is to find the correct candidate moves with an algorithm which isn't brute-force.

[@gurugeorge]

Yeah, that's what I'm thinking - ultimately it's pitting the professors' Chess acumen (and the Chess knowledge/ability of whoever they consult for knowledge, or whatever books and learned articles on Chess they consult) *as embodied in their programming,* versus Kasparov, with a slight advantage to the computer because of its speed of calculating multiple possibilities, plus the absence of any psychological element on the computer's side. Compare and contrast a "baby" AI learning Chess on its own, figuring out its own tactics, etc., etc. - _then_ that would be something like a real test between computer intelligence and person (then the person would be even more likely to lose, I think). I don't know, maybe Deep Blue is already doing something like this, and it's knitting its own internal programming, going beyond what's programmed into it?

[@alexandersapir9079]

Unfortunately, a computer program cannot change its own code (and will never be able to do that). Alan Turing has proved it mathematically. Computers can't program themselves because it's mathematically impossible :(

[@samcavanagh7993]

Alexander Sapir what are you talking about? there are already programs that do just that

[@vijayaraghavanrajagopal475]

that day was incredible

[@paulschwarber3351]

Why this video loads a lot

[@thesupersteve1000]

The computer was controlled by a human as well.

[@Safwan.Hossain]

Kasparov was the best human player so I don't think it would benefit a lot to have it being controlled by a human

[@wingzero2348]

@@Safwan.Hossain Yes it does, did you even watch the documentary. Kasparov was playing against a style that a computer would play but with human interference it changes the whole dynamic.

[@bazza2974]

Ibm cheated

[@sapien153]

How can you cheat in a game where you are allowed to do anything.. lolz

[@kadayori12]

By what? Consulting a human?

[@dominusdone5023]

@@kadayori12 Think of it like this. Youve played a previous match with Kasprov and then you look at how he plays the unexpected move then you modify your computer to play the unexpected and to guess what hes going to do.

[@Mr.Mister420]

This was back in 86 87 .. Now there is no Scope of Beating an engine in Chess

[@x_isaka]

full video?

[@thinkagain2229]

Kasprov is a legeng

[@juanok2775]

To be honest chess is very simple for deep learning but now cleaning your house or cooking a dish oh hell that shit is hard

[@clash5j]

It wasn't that big a deal for me. It wasn't a triumph of machine over human, it was a triumph for the HUMANS that developed Deep Blue

[@paulblart7378]

Yeah but it was a huge milestone for AI research, which was the whole point of building Deep Blue.

[@capnrob97]

And today, 1n 2023, Leela Chess Zero running on $99 Jetson Nano would wipe Deep Blue off the board.

[@sherlyvthomas6095]

Wait so if humanity won first and then AI won the second time then there is still a third match we can still win?

[@decentish8546]

No, because as soon as the game was done IBM dismantled deepblue and destroyed their logs. They were scared of a rematch and also likely cheated in this match.

[@shoeofobama6091]

back then there was a pretty real chance kasparov could have won overall against ibm but now we have chess ais that we can say with confidence that no human who ever lived or ever will live can beat. and artificial intelligence is still in its infancy

[@kekwnet]

@Decentish well that doesn't matter anymore, no human can beat the stockfish engine

[@rickybobby8563]

It doesn't sound like you guys have all the facts correct.. My dad was working at IBM during this period of time so I was paying close attention..

[@gothia6515]

Kasparov is one of the greatest chess players of all time. However, it baffles me a bit that they were surprised by his behaviour. He already proved multiple times that he is a sore loser. He even admited this himself later.

[@lazysuicidal]

I came here from Dr. Stone...

[@phuongpham2019]

greateast man vs greatest group ( technology ) 1 time win but in next time...

[@RonPaul42069]

0:00 Mike Stoklasa?

[@FM-dm8xj]

How ironic, Gary lost because he was too emotional, to intimitated and therefore too human.....

[@primetime3422]

Did deep blue ever play deep blue?

[@unlydoors4u]

Dayum

[@chandanpanda9035]

Who all came from Samay's Stream. Garry is such a legend.🔥

[@riverfrontdetroit1]

This AI is dangerous

[@DarkWizardGG]

More dangerous than ChatGPT?! lol😁😄😅🤖🤖🤖🤖

[@gamegineer11]

I'm from 2076, the robot has taken over the world.

[@robertbudtxu]

Then it took on Kasperov

[@alexmartinez3683]

This was more political than it was to see how good humans are

[@NotGoodAtCombat]

This is scary.

[@balokrumput9074]

Dude AI's scary

[@bikpineda9127]

imagine if he battled stockfish or alphazero

[@DarkWizardGG]

He can fight those AIs toe-to-toe but cant still beat it.😁🤖🤖🤖

[@ShrutiGupta-n3p]

I am

[@armando2814]

es, obviously. The point was to show that a computer, by itself, could beat a human at chess. So if an IBM chess player made some moves for the computer manually in real time, that would be a human beating a human, don't you think? There doesn't seem to be any evidence that Deep Blue was assisted by a human, but it was still a really sketchy match (that for some would be considered cheating) on IBM's part: 1. the engineers (i have no idea why Kasparov's team was ok with this) were allowed to change Deep Blue's code in-between matches. 2. Kasparov's team and IBM had agreed that in order for Kasparov to prepare for the new engine's style of play, he would be allowed to analyze all of Deep Blue's previous publicly played games. When they asked for the games, IBM responded with "well Deep Blue hasn't really played any public games yet so we aren't giving you anything". 3. During the match, Deep Blue played some moves that were supposed to be impossible to execute by a computer of that era. 4. After the match, Kasparov accused IBM of cheating, and demanded they produce proof of their software running the winning moves, but IBM refused. They later quickly dismantled Deep Blue and tried leaving everything behind them (Kasparov even tried to offer a rematch, but they refused). They eventually released Deep Blue's log files, but only 3 years later, and those could have easily been fabricated. Oh one more thing, IBM eventually also open-sourced the code of Deep Thought, Deep Blue's predecessor, and it was found to be unable to perform the crazy moves Deep Blue pulled off.

[@philippemariani9035]

Kasparov won

[@rosnelianadamanik1839]

I-I cant...

[@prashanth.g1945]

then the humans created this algorith are the most intelligent than the world champions themselves! woah, isn't it mindblowing?

[@jorgeespinola944]

I'm here because of Dr. Stone

[@remikamal2302]

Effortless. The same as running a console.log, or lightning up a lightbulb.

[@cashprinter5000]

From beating a grandmaster to fooling millions.....AI sure have gone a very long way

[@PistolPete1998]

Yeah now computers can destroy them now

[@clappedoutmotor]

This is really sad.

[@CezarCezyy]

dont meter what his move Gari Kasparov the al is have good atack end distroy defense , is lose beacouse is not have the time travel deep blue is move very strong end you know what happeneed

[@Insight-n1j]

👍👍

[@حدازوين]

Computers are much better now

[@paulblart7378]

you don't say

[@zyno110]

Yeee

[@oshiezz]

"BBC is a british broadcast service." *Barbecue*

[@Ostiarii]

He is Armenian, not Russian.

[@Unpug]

D:

[@szewei85]

Haha early days of AI yo😄😆

[@WCiacho]

Though, programmers refused to show logs from this game to Gasparow, because there was a bug in a program that made the move unpredictable and senseless, so Gasparow defeat was rather a matter of concidence than intentional action of a program.

[@sincerely-t6t]

if computer can beat human in a chess game, which involves conscious decision and complex brain processes, then AI is more than real.

[@professorx3060]

Playing chess for AI is not a conscious decision though. It's more like a statistical analysis.

[@alberteinsteinthejew]

I bet today a simple chess app can beat Kasparov

[@legionofyuri]

Most likely, considering you can get stockfish on mobile now.

[@BoleDaPole]

the day chess died.

[@sayandas3997]

This shows us just how pathetic we are compared to our very own creation

[@kendal9602]

That's 4 versus kasparov 🤣🤣🤣

[@zarrow50]

chess is not intellectual

[@vthhdg7tidt236]

Who else came here from dtrh

[@danieldbdb]

Unfortunately ibm cheated on this match

[@Avery-df9vy]

how?

[@danieldbdb]

@@Avery-df9vy There was allegedly some "human moves" in there. And they didn't release the prints as promised. Garry adapted his play to face a machine and that move in the ruy lopez was clearly human back then. Ruined his win. What were they afraid of? That was dishonest. They used Garry in their propaganda. Humans would lose to machines, but it shouldn't have been there.

[@Wingly113]

imagine if AlphaChess play against Kasparov, AlphaChess would win by 100 to 0

[@ace942]

Today IBM would be banned for using a computer to make its moves on a chess server.

[@thekaiser4333]

Blah blah blah blah blah blah blah blah blah blah. Please make a proper program about this. Thank you very much.

[@slapmyfunkybass]

Would have been even better if IBM never cheated.

[@sapien153]

How can you cheat in a game where anything goes.. lol

[@godslaya4563]

Hundred years later the sentinel robots will look back at this video as their 1st win over the foolish humankind.

[@alexandersapir9079]

Humans wrote Deep Blue, not robots

[@OldConorWasGreat]

and father*. Most people tend to forget that it needs both genders for human to breed.

[@ultraboombean]

deep blue will be a hero. haahah

[@VyasAnand]

god slaya we don’t even understand consciousness, this is brute force

[@carlhyldborglundstrm9807]

Alexander Sapir We made its life seed. Now it’s making itself.

[@ilferrari]

Kasparov played below his strength in this match. He was trying to second-guess the computer all the time, play anti-computer moves. In 2003 he played against two far stronger computers, and drew both matches.

[@carlfletcherjunior9076]

EXACTLY in my eyes honestly it is more of a draw since they literally had so much research and support against a single talented man

[@Guztav1337]

Considering it was 2½ - 2½ I would also say this was actually a draw. But we have come a long way since then. Good job either way.

[@AntonnyTan]

Thank you sir Kasparov, to returned honor back to humankind. Yet, this algorithm and such would always remind people to the day they defeat us and tend to make us forget, Kasparov revenge.

[@pgreg8528]

Part of chess is figuring Out the opponents strategy. He failed and figuring af that and lost.

[@shuozhang6947]

​@@pgreg8528 The problem was that the computer sometimes had programming errors and made moves that made no sense. However, Kasparov couldn't tell if it was seeing something he couldn’t or making an error. He wasnt able to look at Deep Blue's previous games either, while deep blue analyzed tons of kasparov's games. So kasparov never knew what moves Deep Blue might play. Kasparov was always at a major disadvantage

[@proton_sponge]

No one remembers that Garry Won the first game

[@TechTehScience]

That's because it's not that impressive. One of the greatest Grandmasters beat a chess computer when they were still in their infancy. Garry had far tougher opponents than Deepblue in its first game. This however was big as it marked the beginning of a rapidly developing technology.

[@yungboomer6467]

@@TechTehScience it is impressive

[@pgreg8528]

Both Times the Team didn't have enough time to finish deep blue properly

[@truenoae8689]

@@pgreg8528 dude they had years and years

[@SaintcantfeintIIOKII]

@@TechTehScience noooo he is quoting the punchline from Ted talk of kasporav :/

[@jasonsuriadi2246]

Deep blue: i saw 14,000,605 of our potential futures Winning scenario: 1

[@juretalic333]

nice one haha

[@happygood18]

Got the reference XD

[@armywashere]

I feel the psychological aspect of playing something you are not familiar with, played an instrumental role in Kasparov's defeat. Humans look at "tape" to get a better understanding of their player, but he was not able to do this. Already this put's him at a disadvantage. Him coming up with ideas, such as "anti-computer moves" so quickly was a testament to his genius, his ability to adapt at a dime, exploiting bugs in a computer system to ultimately lead to a win. This is a guy who would go to a dimly lit library with a journal to record good chess moves on a shriveled up piece of paper. I mean it really is not bad at all. He was going against a billion dollar company with hundreds of geniuses and all the resources in the world, all working together to try and beat this one man, and you know what? He still put up a fight.

[@xancarter3037]

when you put it like that it’s actually pretty badass

[@nikita228Rus]

Waaaaaait a minute. Garry win 4 rounds, deep blue only 2. Garry win in this fight, no?

[@davidhiggins330]

That was the first match in Philadelphia. This was a year later and Kasparov narrowly lost the rematch

[@Lemonsieur-m4m]

The sales rep says otherwise.

[@Rogue_Console]

I just went from watching a UFC fight to this...

[@therushinguy1552]

Shabir Ali haha literally just did the same

[@shukurimin5865]

Heheheh me too

[@djquiksilva]

No joke me 3

[@DubstepFan38]

Likewise

[@leetorry]

heh same

[@ASLUHLUHC3]

A lot of salty humans in the comments

[@mrrp405]

I think it must've been incredibly stressful for Kasparov, and he was treated unfairly by the organizers, but even so the achievement DeepBlue accomplished was incredibly.

[@jasongoodacre]

Gary Kasparov lost because he made an absolute insane move early on attacking the knight. This allowed the sacrifice with potential traps opening up the position. He incorrectly assumed that a computer would not play that move.

[@samcavanagh7993]

i'd like to see you play deep blue then

[@Cnut_the_grape]

Yeha that's what we call an inaccuracy and the modern computers call the worst move in chess history. Fuck you stockfish for criticizing every move I play.

[@keikeikei55]

Bots before: managed to win a match against the grandmaster of chess Bots now: BE HONEST, DO YALL THINK IM CUTE OR NAH

[@sunilkc5927]

1 man vs the whole company

[@Ray-mw1fx]

Pewds vs Tseries

[@khein2204]

but to be fair, this whole company could win againts the whole grand masters around the world