Kasparov taps So to challenge Carlsen

An interesting interview with the 13th World Champion was published today titled Garry Kasparov on AI, Chess, and the Future of Creativity, featuring a great deal of chess-related conversation. Notable topics include Deep Blue — the IBM computer which defeated Kasparov in the historic match 20 years ago this month, whether Magnus Carlsen could win a single game in a match against top engines today, if the King's Gambit is underrated, and who is the most likely challenger to Carlsen for the world title. We selected a few chess highlights from the chat with Tyler Cowen for the Mercatus Center at George Mason University. Cowen asked incisive, thoughtful questions and is obviously quite knowledgeable about chess himself.

Kasparov with Tyler Cowen | photo: Mercatus Center

A complete transcript is available via their blog on Medium. The full conversation is well worth a look.

KASPAROV: What we discovered in 1997, when Deep Blue was successful in the match against a current world champion — and I was world champion at that time — that chess could be crunched by brute force once hardware got fast enough, databases got big enough, algorithms got smart enough.

But even while Deep Blue was victorious, it was anything but intelligent. Of course, you can start arguing about the definition of intelligence because, by the definition of its output, the grandmaster-level chess, Deep Blue was intelligent, at this incredible speed. At certain points, Deep Blue could reach a phenomenal speed — even today it’s phenomenal, but 20 years ago it was mind-boggling — of 200 million positions per second. It offered us very little dream of insight into the mysteries of human intelligence.

COWEN: You think complacency is the main human bias you see in top chess games?

KASPAROV: It’s complacency if you are the winning side, but it could be also the desperation if you are the losing side. One of the greatest problems against any machine is how to deliver the final blow to win the game. Even if you lucky and you get a very good position, then again, you have to play with precision that is not required in human chess...I would say that if we have Magnus Carlsen, the current world champion, offered an opportunity to play against the computer, say 10 games match, on the conditions that he wins the match if he wins one game, which means he can play all the games without thinking about losing the game by making a mistake. He can play without this terrible psychological pressure, maybe there’s still a chance.

COWEN: You’ve been a pioneer in what’s sometimes called advanced chess, freestyle chess, or centaur chess, where you pair a human being with a computer or a set of programs. Today, 2017, do you still think it’s the case that a human paired with a set of programs is better than playing against just the single strongest computer program in chess?

KASPAROV: I think so. Again, it depends on the qualification of the operator.

COWEN: Sure, if it’s the best operator in the world, whoever that may be. Maybe yourself, maybe Anson Williams.

KASPAROV: By the way, I exclude myself from this category because I’m not a very good operator. I’m a very good chess player. A great operator does not have to be necessarily a very strong player.

COWEN: What makes for a great operator?

KASPAROV: Someone who can work out the most effective combination, bringing together human and machine skills. I reached the formulation that a weak human player plus machine plus a better process is superior, not only to a very powerful machine, but most remarkably, to a strong human player plus machine plus an inferior process.

At the end of the day, it’s about interface. Creating an interface that will help us to coach machine towards more useful intelligence will be the right step forward. I’m a great believer that, if we put together a good operator — still a decent chess player, not necessarily a very strong chess player — running two, three machines and finding the best way to translate this knowledge into quality moves against Rybka Cluster, I would probably bet on the human plus machine.

You can also listen to the full podcast audio:

COWEN: The chess opening, the King’s Gambit, 2.f4. Underrated or overrated?

KASPAROV: It’s a great piece of history, but it has no relevance today...I look at these openings, and I just come to the conclusion that they simply don’t stand modern technique of defense and counterattack. They’re not feasible. If you are looking for a good game, professional game, they’re not feasible.

By the way, I was the one who reinvented and brought back the Scotch, the Evans Gambit. But I have to realize that . . . OK, I don’t play anymore, but the way chess has developed made most of these openings — not all of them, mainly gambits — simply not feasible weapons at the top level.

I don’t think that the aptitude of playing chess is anything but aptitude for playing chess.

COWEN: We both know Ken Rogoff, highly successful as an economist and a very good chess player. If we take most top chess players at the very top, and somehow they could not have played chess — in other endeavors, how successful do you think they would have been?

KASPAROV: I don’t think that the aptitude of playing chess is anything but aptitude for playing chess. There’s no universal rule. Some people could move to other areas.

I was very, very lucky that I grew up with my mother, who helped me to get education outside of chess. I always had various interests and looked for life in its entirety. That’s why after I stopped playing professionally in 2005, then already with the help of my wife, I could move from chess to other areas. Now I have a very full life without playing chess professionally. But that’s not the rule. That’s not happening with other players. Some of them, I’m sure, could do extremely well.

COWEN: Perhaps you may recall your 1999 game against Topalov. At a key moment in that game, you play Rxd4. When you played Rxd4 and made that sacrifice, did you see at the time the later brilliant moves coming up: your c3, your Bf1, your Rd7? Or did you just feel there’s going to be something that comes up? How much was it one or the other?

KASPAROV: I’m not sure exactly at what point I saw the final combination. First of all, after playing Rxd4, I knew if he would have taken the rook, which was a mistake . . .

COWEN: Right, he should have played Kb6.

KASPAROV: Exactly, and he’d get very good position, but I knew that is already good for me, though I was not sure about the outcome immediately. But within a move or two, I saw the whole line. I was quite lucky, but at the same time, it prevented me from finding the forced win in the middle of the combination because at one point, I made a second-rate move just because I already saw the whole line.

COWEN: It was so aesthetically pleasing.

KASPAROV: I just looked for obvious moves for black. By the way, both me and Topalov missed a much better defense for black that could have prevented this combination, though for the endgame . . .

COWEN: You mean 30… Rhe8? Does black hold with that?

KASPAROV: Yes, yes. The endgame, it’s much better for white, though it’s not winning. The move pointed out by Kavalek, Ra7 immediately, it would be winning by force. It would be a mating net from where black king couldn’t escape. Because I saw this, it’s all the way up to the Rd7.

You look at the time I spent, I played, almost blitzed because I was so anxious to reach the position and to demonstrate how far I could see it. Also, I was amazed by the beauty of this geometry.

COWEN: I didn’t prepare you for this question. You simply remember, correct? The game against Topalov, I didn’t come to you before this session and tell you to go back and look at the game. You simply remember.

KASPAROV: I have to confess that I don’t remember the exact number of the moves. I can reconstruct the game, but when I said Rh8 . . .

COWEN: That’s when he played Qc4?

KASPAROV: Of course, exactly. Then it would end up with the endgame, when I would be changed down, but I had rook and bishop and several pawns for his two rooks. If I remember correctly, probably it’s winning, but it’s not forced.

COWEN: Would a top chess engine have played Rxd4 today?
[Editor's note: Yes, and you can find the game in our database to explore it in depth.]

KASPAROV: I didn’t check, it depends on the relation of Kb6 because even without calculating the whole combination, they will see immediately that white could force a draw.

By the way, that’s what Topalov thought, that I wanted to make a draw. White could force a draw while, after Kb6, every engine shows immediately that black is better. I mean slightly better. But the evaluation, it’s clearly in black’s favor, so that’s why they wouldn’t take the rook. I think at that time, white had no choice but to take on d4 because otherwise, the whole position . . .

COWEN: Then black has much more space, yeah.

KASPAROV: Exactly, exactly.

COWEN: Who is the most likely challenger to Magnus Carlsen this next time around?

KASPAROV: Considering the logic of the world championship history, Magnus should face opposition from a younger player, or the same age. So I would say there are three players that, I guess, could challenge him. Maxime Vachier-Lagrave, I would rate him as least probable among the three. Second, Caruana, and the most likely, Wesley So.

The reason I put Wesley ahead of two others is that he has phenomenal concentration, absolutely phenomenal, and that’s very important. These days, concentration trumps everything. I would put it on top of other things, though he’s a very, very good player. I think that he might be the most dangerous for Magnus.

COWEN: Because he has nothing to lose, he feels?

KASPAROV: No, because his concentration could overboard Magnus. Magnus’s concentration was always his strength. He was always very concentrated. He could mobilize all his resources for the game. We saw Magnus could feel...despite again I guess So has to qualify for the Candidates; he still has to win it.

But against Wesley So, Magnus could have some serious problems in reading his opponent, and also messing up his energy. So could reflect. I think his match with Caruana could be also uncertain, and if Vachier-Lagrave makes some improvements, more psychological improvements, he could be also dangerous. He’s the same age; Caruana is younger. So is even younger than Caruana.

So I would bet on these three as one of the Magnus challengers, ranking them in the following order: So, Caruana, MVL.