From Garry Kasparov's MasterClass

Computers and Chess

The sting of Garry’s loss to Deep Blue has softened over the years. He looks back on that defeat and forward to the benefits of computers in chess training.

Topics include: Playing Deep Blue • Computers Limiting Creativity


The sting of Garry’s loss to Deep Blue has softened over the years. He looks back on that defeat and forward to the benefits of computers in chess training.

Topics include: Playing Deep Blue • Computers Limiting Creativity

Garry Kasparov

Teaches Chess

Learn More


Deep Blue was as intelligent as your alarm clock. Though of course, losing to 10 million dollar alarm clock was not the most pleasant experience. But there was a general belief among computer scientists that chess could serve as an ultimate test for computer's intelligence, artificial intelligence. That was not the right approach. Because machines could conquer the game of chess by using the brute force of calculation. It was an important experiment. I won the first match. I was not successful in the second match. And IBM walked away, dismantled the machine. I understand the rationale of them doing it, because they had already won. They couldn't prove the result. I have to confess that my preparation for the second match was lousy. I didn't expect-- again, I have to give credit to IBM's scientific team. I didn't expect so much progress that they could make. Though of course, the conditions of the match will not 100% fair to a human player. I was under constant pressure, since IBM was not only player, but also an organizer. While analyzing these games 20 years later, I could realize that the quality of the match was not perfect-- far from perfect. For anybody who wants to realize how much progress machines have made since 1997, you can look at these games. And you will find out that some moves that were celebrated by chess world in 1997, Deep Blue moves, as a genius defense like in game five, these moves will be harshly criticized by chess engines today, correctly stating that I missed some of the big chances. And again, there were mistakes on both sides. But it was an important event since it moved forward man versus machine competition. The problem of human chess while facing a silicon monster is not that chess could be mathematically solved. That's one of the great misunderstandings. Chess is a mathematically infinite game. When I say mathematically infinite, it's probably an exaggeration. But when we talk about numbers like 10 power 45, which is more than number of atoms in the solar system, that's, let's agree, it's infinite. While we are talking about machines conquering the game of chess, probably it takes us to the wrong path, thinking about solving the game while we are talking about machines being superior to the best human players. And I can tell you with absolute confidence in a game between two humans, two best human players at the world championship match, there is nothing like perfect chess, perfect game. Some games are really good. But you can always find the tiny, tiny inaccuracy. I'm not talking about blunders or bad mistakes-- inaccuracies. A very good game played by a top player, say the world champion, say it's 50 moves-- it could be 45 good, solid moves, four great moves, but most likely it will be one tiny inaccuracy, which makes no difference when you face another human. The problem is, when you face t...

Elevate your game

At age 22, Garry Kasparov became the youngest world chess champion. After beating Bobby Fischer’s peak rating, he outranked his fiercest competitors for over twenty years. Now, Garry is ready to share the chess strategy that made him a six-time world champion. Through detailed lessons, including his favorite openings and advanced tactics, you will develop the instincts and philosophy to become a stronger player.


Students give MasterClass an average rating of 4.7 out of 5 stars.

this has been a great experience, kasparov is the legend and he has inspired me since i was little. my dream is to become a grandmaster and hopefully play kasparov one day.

First, thank you, Garry, by the brilliant classes!! My chess became better after those masterclass. There were improvement in all fields of chess (tactics, strategy, opening and endings). I'm so grateful for that. I hope Garry returns to another Masterclass. Who knows? rs

Wow. I grew up playing chess with my dad. Wish I would have had this then.

I loved the way how Garry explained chess. The best is to mix inspiration and actual tactics, strategy. Well done!


Larange T.

I got revenge for Kasparov by beating Stockfish 8(real!!!) Hope you like it... 1.d4 d5 2.c4 dxc4 3.e3 e5 4.Bxc4 exd4 5.exd4 Nc6 6.Nf3 Bb4+ 7.Nc3 Nf6 8.O-O O-O 9.h3 h6 10.Bf4 Bf5 11.Re1 Na5 12.Bf1 Be6 13.a3 Bxc3 14.bxc3 Rc8 15.Qc2 Bd5 16.Nd2 c5 17.dxc5 Rxc5 18.Rad1 Rc8 19.Be3 Re8 20.Qb2 Qc7 21.Bd4 b6 22.c4 Ba8 23.c5 h5 24.cxb6 axb6 25.Bxf6 Rxe1 26.Rxe1 gxf6 27.Qxf6 Qd8 28.Qe5 Rc2 29.Re3 Rxd2 30.Rg3+ Kf8 31.Qh8+ Ke7 32.Rg8 Nc6 33.Rxd8 Nxd8 34.Qc3 Rd1 35.Qc7+ Kf8 36.Qxb6 Bc6 37.Qc5+ Kg7 38.Qg5+ Kf8 39.Qxh5 Rd5 40.Qh6+ Ke8 41.Qh8+ Ke7 42.Qc3 Ba8 43.a4 Ne6 44.a5 Kd7 45.a6 Kd6 46.a7 Rd4 47.Qc8 Bc6 48.Qxc6+ Kxc6 49.a8=Q+ Kd7 50.Qa7+ Ke8 51.Qa8+ Kd7 52.Qa3 Rd6 53.Bb5+ Ke7 54.g4 Nf8 55.f4 Nd7 56.f5 f6 57.h4 Nf8 58.h5 Ne6 59.fxe6 f5 60.Qxd6+ Kxd6 61.h6 Kxe6 62.h7 fxg4 63.h8=Q Kd5 64.Qc3 Ke6 65.Qd4 Ke7 66.Qd5 Kf6 67.Qe4 Kf7 68.Qe5 g3 69.Kg2 Kg6 70.Qf4 Kg7 71.Qf5 Kh6 72.Qg4 Kh7 73.Bc4 Kh6 74.Bf7 Kh7 75.Qg6+ Kh8 76.Qg8# 1-0

Lewis G.

This goes down an interesting path: the computer as a training partner. We can look for confirmation and advice, and learn with the help of the machine. We learn to refine our own creativity and imagination on the board.

Gary P.

ohh boy Computers, Fischer, the end of chess i have heard it all and they are all mad men. Humans will never master chess That is the beauty of it. Fischer said it was all just memory and chess was dead. He was the greatest candidate player and worst world champion a real shame he was scared to play as champion then his way out was to belittle the very game that made him a genius. Don't bite the hand that feeds you. Chess is great. Once you break or take something apart you spoil its beauty. enjoy it for what it is.

Michael M.

I imagine that middle section -- spend time analyzing your own games and especially the mistakes they contain --would be good advice for musicians, actors, CEO's, race car drivers, and just about anybody who does anything requiring talent or skill. Very good food for thought even outside of chess, but certainly good advice for our chess games.

Aaron B.


Thomas M.

I've used Computers for chess for while, but I've failed to learn how to use them for Analysis.

Krishna P.

To think that Computers are stronger , in intelligence, than humans is like saying that Cranes are superior to humans because they can lift more weight than humans. Both are incomparable.

David B.

AI program learns chess from first principles, no opening book, no end game tables, no training set of replayed games, is able to beat stockfish within 24 hours of self play!!! - learns entirely on its own, amazing.... see the discussion on its changing selection of openings.... 8 predict it will soon present a whole new view on all aspects of chess play.

John C.

Some interesting points here. Garry notes that artificial intelligence in the context of chess is effectively brute force, at least it certainly was until recently. There are systems out there now that are self-learning, often without any knowledge of a game initially (I think the recent Go AI is one of them, alongside ones that learn to play arcade games), and this might inject a little creativity in the future. Now, the point about the psychological side is very interesting here: computers don't have an emotional response as yet. They don't get cocky, they don't worry about a move they made earlier, they don't wonder if maybe they shouldn't have had that burrito before the game, or get irritated with the opponent's breathing. It would be an interesting development if we ever reach a stage where chess computers develop a personality and emotional response - although their dispassionate analysis is a useful tool for the moment. Eric's comment about modern players not trusting themselves is a good one. Although Garry's comment about Magnus using the computer for confirmation, not the definitive answer, was intended to be a positive one, it did seem interesting that the confirmation still needs to be done by a top player. However, we come back to the dispassionate analysis and brute force on offer. It seems like blitz and bullet chess is progressing as a reaction to the slow, careful, computer-assisted analysis perhaps?

David B.

I have a grandson who learned chess by playing on the Nickelodeon online chess game. He plays at almost blitz speed, I try to advise him, but he's 5 moves past me before I voice my concern.... the computer is addicting.... Fritz 15 on the PC and Chess Genius on my kindle work for me. I've found suggested moves on chess genius are not good, so I don't trust it on analysis, it plays well against me otherwise (?). Annotating games is still a problem as many programs don't display comments or allow editing. I'm still pretty much old school, and have a small collection of real chess sets.