I have seen a lot of debate about who is better between Carlsen and Kasparov.

Could a match between Carlsen and the exact same Deep Blue configuration that beat Kasparov be used to settle this?


I'd say no. Playing against a human being is a lot different than playing computers. Here is an article (unfortunately, it is in Dutch) about a German player with a rating under 2100 who regularly beat the top chess programs, by playing strange openings and long-term sacrifices. This was a few years after the Kasparov - Deep Blue match.


Deep blue was dismantled soon after playing Kasparov.

So, I guess, we will never know what would have happened. However, if it does happen, then I would favour Carlsen winning against the computer. At the time that the match against Kasparov was played, computers weren't really stronger than humans.

And to be honest, I think, Kasparov lost because of his usually attacking instincts and the fact that he played openings that he never usually played.(This was because, he feared the computer had access to his games and he wanted to take the computer out of theory known at that time.) Even then, Deep Blue barely managed to win when Kasparov fell into a known trap in the last game.

So, I would say if this hypothetical match did happen and Carlsen plays at his best, Carlsen would easily win against Deep Blue.

However, this does not mean that in the current times, humans would beat computers without odds. Deep Blue was far weaker than modern-day engines, say Stockfish or Komodo. If you want to see the strongest human chess player vs strongest engine and who will win, then, engines would reign supreme.

  • In a recent interview Carlsen said he is not going to play any computer any time soon: youtube.com/watch?v=Drxz3PH9gJE. Also "However, if it does happen, then I would favour Carlsen winning against the computer": false; nowadays computers (even on your laptop) are (at full depth) way over any GM level (around ELO 3300 at max, although ELO comparisons for computers are to be taken in quotation marks), thus I wouldn't be so sure Carlsen (or any other) could win easily, he himself (minute 4:42) says "I wouldn't have any chance at all against the best computer".
    – gented
    Dec 19 '16 at 9:32
  • The human and engine elo scales aren't they well calibrated, but 3300 is an extremely low estimate. 3600 would be a lot closer. Jul 27 at 16:56

Chess engines and hardware since Kasparov played Deep Blue continue to improve.

Recently (Dec 2016) Stockfish achieved a rating of 3478: http://www.computerchess.org.uk/ccrl/404/

Carlsen fares reasonably well against it (on a reasonably normal computer) though its a strong engine.

Carlson playing against Deep Blue also has a couple of advantages Kasparov didn't have. Carlsen himself can use computers to prepare to play a computer AND Deep Blue has now has games on record so can have its own play analyzed.

So a Carlsen/Deep Blue game would tell us little. More interesting would be a Carlsen game against a modern chess engine with less time restrictions per move on searches, and on optimized hardware beyond what you can get at home.

Even then advantage goes to machine (likely).

Also check out discussion here


I would say no.

In an interview with GM Illescas, who was on the IBM team, I remember that he said Kasparov complained Deep Blue was not designed to beat a GM, it was designed to specifically beat him.

As the question says with the exact same configuration, I would say no because Deep Blue didn't include Carlsen’s games in its database in 1996 for an obvious reason.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.