Catch up on stories from the past week (and beyond) at the Slashdot story archive

 



Forgot your password?
typodupeerror
×
Classic Games (Games) AI

Magnus Carlsen Wins 8th World Chess Championship. What Makes Him So Great? (espn.com) 42

"On Friday, needing just one point against Ian Nepomniachtchi to defend his world champion status, Magnus Carlsen closed the match out with three games to spare, 7.5-3.5," ESPN reports. "He's been the No 1 chess player in the world for a decade now...

"In a technologically flat, AI-powered chess world where preparation among the best players can be almost equal, what really makes one guy stand out with his dominance and genius for this long...? American Grandmaster and chess commentator Robert Hess describes Carlsen as the "hardest worker you'll find" both at the board and in preparation. "He is second-to-none at evading common theoretical lines and prefers to outplay his opponents in positions where both players must rely on their understanding of the current dynamics," Hess says...

At the start of this year, news emerged of Nepomniachtchi and his team having access to a supercomputer cluster, Zhores, from the Moscow-based Skolkovo Institute of Science and Technology. He was using it for his Candidates tournament preparation, a tournament he went on to win. He gained the challenger status for the World Championship and the Zhores supercomputer reportedly continued to be a mainstay in his team. Zhores was specifically designed to solve problems in machine learning and data-based modeling with a capacity of one Petaflop per second.... Players use computers and open-source AI engines to analyze openings, bolster preparation, scour for a bank of new ideas and to go down lines that the other is unlikely to have explored.

The tiny detail though is, that against Carlsen, it may not be enough. He has the notoriety of drawing opponents into obscure positions, hurling them out of preparation and into the deep end, often leading to a complex struggle. Whether you have the fastest supercomputer on your team then becomes almost irrelevant. It comes down to a battle of intuition, tactics and staying power, human to human. In such scenarios, almost always, Carlsen comes out on top. "[Nepomniachtchi] couldn't show his best chess...it's a pity for the excitement of the match," he said later, "I think that's what happens when you get into difficult situations...all the preparation doesn't necessarily help you if you can't cope in the moment...."

Soon after his win on Friday, Carlsen announced he'd be "celebrating" by playing the World Rapid and Blitz Championships in Warsaw, a fortnight from now. He presently holds both those titles...

The article also remembers what happened in 2018 when Carlsen was asked to name his favorite chess player from the past. Carlsen's answer?

"Probably myself, like, three or four years ago."
This discussion has been archived. No new comments can be posted.

Magnus Carlsen Wins 8th World Chess Championship. What Makes Him So Great?

Comments Filter:
  • by fahrbot-bot ( 874524 ) on Saturday December 11, 2021 @05:40PM (#62070651)

    Magnus Carlsen Wins 8th World Championship. What Makes Him So Great?

    Um... His name is "Magnus" [wikipedia.org]:

    Magnus, meaning "Great" in Latin, ...

    • Re:Raises hand ... (Score:4, Insightful)

      by ShanghaiBill ( 739463 ) on Saturday December 11, 2021 @06:04PM (#62070709)

      Magnus reflects the influence of AI on chess. In the past, championships had dramatic games with aggressive attacks. But chess engines like AlphaZero don't play that way. They grind out victories by slowly amassing and exploiting tiny positional advantages.

      Magnus has adopted a similar style. An example of this new dominant style was game six of the just-completed match. Early in the game, Magnus gained a slight advantage by exchanging his queen for both of Ian's rooks. Then he slowly and methodically exploited that tiny advantage in an exhausting 136 move game lasting over seven hours. No other championship game has ever lasted as long.

      AI has changed chess forever. There is no going back.

      • by Anonymous Coward
        Magnus also reflects the influence of AI on chess by virtue of his opponents relying so heavily on AI-based research that their abilities to play unknown positions haven't kept up. And certainly haven't kept up with Magnus'.
        • his opponents relying so heavily on AI-based research that their abilities to play unknown positions haven't kept up.

          This is backward. Playing against an AI engine gives a player more opportunities to explore new positions.

          It is easy to set up a board position and let an engine crunch through the options and give an analysis of each path. You can't do that either quickly or reliably with a human partner.

          And certainly haven't kept up with Magnus'.

          This is also backward. Magnus plays with a style similar to the AI engines. So practicing against an engine is better preparation than playing against lower-ranked human opponents.

      • Magnus reflects the influence of AI on chess. In the past, championships had dramatic games with aggressive attacks. But chess engines like AlphaZero don't play that way. They grind out victories by slowly amassing and exploiting tiny positional advantages.

        In general, yes, but Alpha Zero has played some amazing attacking games, even against Stockfish.

        • by nagora ( 177841 )

          Alpha Zero doesn't play chess - it imitates playing chess. It plays chess in much the same way that my car "runs a marathon".

          AlphaZero has no grasp of chess whatsoever.

          • Alpha Zero doesn't play chess - it imitates playing chess.

            Like there is any difference.
            If you win at a game with identical rules to chess, then
            you are playing chess.

            • by narcc ( 412956 )

              I'm going to agree with you for a totally different reason. The way humans play chess at a high level and computers play chess is depressingly similar.

              If you want to become more than just a casual player, you've got to dedicate an absurd amount of time learning what chess people call 'theory'. That doesn't mean learning rules or principles or anything like a normal person might think from the word 'theory', it means memorizing thousands of board positions and what the best thing to do is in each case.

              If

              • by nagora ( 177841 )

                I'm going to agree with you for a totally different reason. The way humans play chess at a high level and computers play chess is depressingly similar.

                AlphaGo does not play chess remotely like a human.

                • by nagora ( 177841 )

                  I'm going to agree with you for a totally different reason. The way humans play chess at a high level and computers play chess is depressingly similar.

                  AlphaGo does not play chess remotely like a human.

                  It plays Go. AlphaZero doesn't play chess the way humans do :)

              • by Budenny ( 888916 )

                No, not true. Theory means opening theory, and no, you don't have to memorize huge chunks of it. If this is your approach you are doing something wrong.

                Yes, to play at grandmaster level you have to know openings in depth, your opponents will, and you will fall into traps otherwise.

                But no, as a club player you don't have to know all this. A modicum of the main variations of what you intend playing, and a good understanding of the typical positions will do fine.

          • "Alpha Zero ...imitates playing chess." So do I. What's your point?

      • Re:Raises hand ... (Score:5, Interesting)

        by phantomfive ( 622387 ) on Sunday December 12, 2021 @12:33AM (#62071333) Journal

        Magnus reflects the influence of AI on chess. . In the past, championships had dramatic games with aggressive attacks. But chess engines like AlphaZero don't play that way. They grind out victories by slowly amassing and exploiting tiny positional advantages.

        That seems extremely unlikely, since Magnus played that way before AlphaZero existed. Also the concept of "slowly amassing and exploiting tiny positional advantages" is an old one, that is how Capablanca and Lasker played. Finally, Peter Svidler and Kasparov have both pointed out that AlphaZero has more of a romantic style, with aggressive attacks, when compared to other computer engines.

        • Magnus won because he's a net-better player. Whether it's a ground-down victory, or deliciously deliberate, the strategies he uses win and draw, but win sufficiently to retain title. That's chess. It's intellectual BDSM.

      • by teg ( 97890 )

        Magnus reflects the influence of AI on chess. In the past, championships had dramatic games with aggressive attacks. But chess engines like AlphaZero don't play that way. They grind out victories by slowly amassing and exploiting tiny positional advantages.

        Magnus has adopted a similar style. An example of this new dominant style was game six of the just-completed match. Early in the game, Magnus gained a slight advantage by exchanging his queen for both of Ian's rooks. Then he slowly and methodically exploited that tiny advantage in an exhausting 136 move game lasting over seven hours. No other championship game has ever lasted as long.

        AI has changed chess forever. There is no going back.

        I disagree. What you are describing - slowly amassing and exploiting tiny positional advantages - is the classical way used by many players, e.g. like Capablanca, Petrosian, and Karpov: Get a small advantage and crush your opponent, slowly.

        What we have learned from computers - especially AIs, but also others - is that they put a bigger premium on piece activity vs. material balance, and even heavier focus on constriction, like a white pawn on h6 in a fianchetto position (KID, GID). Nothing too crazy like t

  • chess player, right?

    Not the number 1 chess player.
    • by splutty ( 43475 )

      Computers don't get ranked. So "human" is superfluous.

      • by ShanghaiBill ( 739463 ) on Saturday December 11, 2021 @07:49PM (#62070859)

        Computers don't get ranked. So "human" is superfluous.

        Likewise, the Olympic champion weightlifter is a human. Forklifts aren't allowed to compete.

        • Re: (Score:1, Insightful)

          by iggymanz ( 596061 )

          Fork lifts aren't entered into competitions with humans, but chess programs and their hardware are. Chess programs are ranked, even by human ranking system such as Elo, and the programs are hundreds of Elo points higher than humans. After over forty years, the chess programs began to beat humans in the mid 1990s, something once said to be impossible. Chess programs are also ranked into comparison with each other.

      • False, computers have been ranked, such as with Elo system, look it up. Computer chess systems are also ranked among each other, look it up.

        • by splutty ( 43475 )

          We're talking about world championships. There are no computers in the world championships, hence they're not ranked in any sense that's relevant to this article.

          The fact they can be assigned an Elo number doesn't mean they're 'ranked' for world championship purposes. There's a reason they have their own list :P

          So yes. Technically you're right, if you want to be pedantic about it. It also has nothing to do with the article, which the comment was about.

  • by seoras ( 147590 ) on Saturday December 11, 2021 @10:34PM (#62071031)

    Game 5 was epic. Both were exhausted after one of the longest final games ever. It ended in a draw.
    From then on Nepomniachtchi crumbled. Made an awful mistake in game 6 allowing his Bishop to get pinned and taken.
    If you watch the video of the game you can see Magnus shaking his head in disbelief after studying it to see where the non-obvious advantage to that move was and realising there was none.
    Game 7 was a draw with Magus playing black but Magnus won game 8 as white and then game 9 as black.
    Game 10 another draw with Magnus winning as black the final game 11 when Nepomniachtchi blundered again.
    Commentators felt Nepomniachtchi had effectively given up or lost interest by the final game.
    Takes nothing away from Magnus and I can't wait to see if he retains both World Rapid and Blitz Championships.

    • by quantaman ( 517394 ) on Sunday December 12, 2021 @12:21AM (#62071289)

      Game 5 was epic. Both were exhausted after one of the longest final games ever. It ended in a draw.
      From then on Nepomniachtchi crumbled. Made an awful mistake in game 6 allowing his Bishop to get pinned and taken.
      If you watch the video of the game you can see Magnus shaking his head in disbelief after studying it to see where the non-obvious advantage to that move was and realising there was none.
      Game 7 was a draw with Magus playing black but Magnus won game 8 as white and then game 9 as black.
      Game 10 another draw with Magnus winning as black the final game 11 when Nepomniachtchi blundered again.
      Commentators felt Nepomniachtchi had effectively given up or lost interest by the final game.
      Takes nothing away from Magnus and I can't wait to see if he retains both World Rapid and Blitz Championships.

      I doubt anyone at that level gives up or loses interest.

      Rather, I suspect Nepomniachtchi simply burnt out.

      A lot of Carlsen's training is based around needing to maintain performance over a long tournament [reddit.com].

      What happened during the world championship? Things were even and then they played an extraordinarily long (and being the world championship, intense) game, and Nepomniachtchi blundered.

      In the subsequent games Nepomniachtchi continued to blunder. How many people here are programmers? You can go deep into a problem for a couple of days, but then you tend to be running on fumes for a bit after that.

      I suspect Nepomniachtchi put everything he had into game 6, and once it was done the effort of that game 6 combined with his preparation prior to the tournament meant that he was all out of gas and subject to lapses in concentration that resulted in blunders.

    • Caruana commented and said that Nepomniachtchi was playing like a 2300 player.

    • by Vlad_the_Inhaler ( 32958 ) on Sunday December 12, 2021 @06:52AM (#62071879)

      Game 5 was epic. Both were exhausted after one of the longest final games ever. It ended in a draw.

      Uh, no.
      Games 1-5 were draws and the longest one (game 2) lasted 58 moves. There was also a day off between games 5 and 6.
      Game 6 was the one which lasted 136 moves and it was not a draw. Either player could have won it in the scramble before the first time control (at 40 moves), the critical sequence happened after move 25 when Carlsen ended picking up both rooks for the loss of his queen. Once the dust settled - the time control - Carlsen deliberately kept things unbalanced and complicated, partly to tire Nepomniachtchi out and partly because one of his major strengths is just this kind of position. Carlsen gradually defused Nepo's attack and kept him on the rack until move 130, when Nepo finally cracked.
      Game 7 was uneventful, both players were exhausted.
      Game 8 saw Nepo trying to break back, but he missed a pawn move which trapped his bishop. Carlsen was now two wins up.
      Next rest day. By now Carlsen was keeping things complicated but not really looking to attack, draws were enough for him.
      Game 9 saw a rested Nepo trying very very hard to break through. Too hard, he missed something simple and lost again.
      Game 10 was a tame draw with one player shattered and the other happy with the situation.
      Game 11 was embarrassing. Nepomniachtchi had to take 3½ points from the last 4 games and he made another blunder to end the torture then and there. At least one of the commentators - Giri - saw it as losing deliberately, no idea if he was right.

      Carlsen likes to avoid his opponent's opening preparation and he has a repertoire of offbeat-but-harmless moves in standard openings, aimed at taking the opponent out of their "book". The idea is that he is the better player and has a good chance of grinding the opponent down.

      • by teg ( 97890 )

        You also misremembered the games slightly - you mixed up 8 and 9. Game 6 was the long one, game 8 (was where Nepo blundered Qa3), game 9 was where Nepo blundered the trapping of his bishop with c5, and finally game 11 was where he blundered with g3.

        • oh yeah, with three blunders - at that level - I forgot which was which.
          With two of them, I only knew it was a blunder when Carlsen exploited them (and the engines flagged them). Amusing was that Judit Polgar missed one of them, the one in Game 9 - she had thought c5 was the best move.

  • by Budenny ( 888916 ) on Sunday December 12, 2021 @05:48AM (#62071787)

    The piece referred to is the usual vague opinionating which people who are either not players at all or not strong enough players put forward about chess.

    The two main things distinguishing Carlsen from his peers are one his positional judgment, two temperament.

    His great temperament asset is that he is stable, doesn't panic or get knocked off balance, and is very pragmatic and realistic. You can see that compared to Nepo in this series, confronted with similar level of challenges, he picked his way through the alternatives much better.

    This is helped by his superior positional judgment. He seems to have a better sense of the balance of advantage in a position. Humans don't think like the computers - they rely on pattern recognition and intuition, so when evaluating they are continually looking at a position without really analyzing and concluding that there are enough of the right elements grouped in the right way that when they get to it, there will be tactical courses for either side to be found. You can hear this in action if you ever listen to grandmasters thinking aloud as they evaluate a position.

    Carlsen of course is capable of accurate calculation and of seeing resources quite some moves into forcing sequences, as all players at his level are. But the key asset is probably the judgment.

    Has AI changed the game? Not that I can see. Look at the matches of the 20c and there is not much difference. The endless QGDs between Capablanca and Alekhine for instance, very closely matched, lots of draws. Recent matches have been between players whose strength is close. There is also an emphasis on conservative play, caused largely by the format. The suggestion that Carlsen is somehow different in his exploitation of small advantages also doesn't survive scrutiny. This is the way chess is played, and always has been for all of the 20c and the present century.

    Carlsen is just that bit better, both in temperament and in judgment, and that's the only remarkable thing about him.

  • The article https://tech.slashdot.org/stor... [slashdot.org] makes the point that people who rely on Google and wikipedia for their "knowledge" don't remember as much. And here we have a chess player who does it all himself beating a chess player who relies on computers and AI.

    • by Budenny ( 888916 )

      A chess program is useful in two cases. First, if you are researching an opening variation, it can give best lines of play that you might otherwise not have hit on.

      Second, it can analyze a position and give what it regards as best lines of play. And as you try out your own ideas against it, it will show best moves against them.

      However, there is one very important proviso that you have to bear in mind while doing this. The computer may, correctly, evaluate a given position as even, or advantage for one si

To be or not to be, that is the bottom line.

Working...