“My biggest advantage is that I am better at chess.”
It’s classic Magnus: blunt, brusque, bordering on cocky, but when you consider his abilities, perhaps just realistic. Yet seen from a different angle, it’s also a profound take on prediction.
To explain why, I need to go back to 1993 and Super Bowl XXVII, Cowboys vs. Bills. I remember watching the pregame show. There were three talking heads and each was going to give their prediction about the upcoming game. There doesn’t seem to be video of the pregame show online, but it went something like this:
Commentator 1: This is the Cowboys’ first ever Super Bowl appearance. Usually it takes two: one to know what it feels like, one to win it all. The Bills have been in the Super Bowl the last two years. They know what it takes to win and they’re ready. I like the Bills.
Commentator 2: The Bills have three receivers with over 500 yards on the season and they overwhelm defenses with the pace of their no-huddle offense. The Cowboys have never faced an offense like this and their secondary isn’t set up to handle it. I’ll take the Bills in a shootout.
The commentators were former coaches or players. Their predictions were based on specific and detailed facts about the two teams. They clearly knew a lot about football and their takes seemed to make sense. But then it was time for the third guy to give his prediction.
Commentator 3: The Cowboys will win because they’re a much better football team.
The prediction seemed glib - much like Carlsen’s comment - but proved to be on the money. The result of the game was that the Cowboys won 52-17, the most lopsided Super Bowl in history. Why did the third commentator get it right when the other two, who seemed to know so much more, got it wrong? Was it just luck?
The first two commentators used the inside view whereas the third commentator used the outside view. These terms were coined by Daniel Kahneman and Amos Tversky to describe two styles of forecasting. The inside view involves making predictions based on specific details. The outside view uses statistics about the general class that the current situation belongs to, or a big-picture view more generally.
Experts on a subject often use the inside view, basing on their predictions on highly specific details that they understand because of their expertise. In some cases this blinds them to the big picture. If your prediction is based on a very specific nuance, you’re implicitly making the argument that apart from the one thing you’re focusing on, all the other details come out equal. But there is no reason this has to be the case: the situation overall could greatly favor one side or the other. In other words, if your prediction is based on a specific, narrow detail, what you’re leaving out is much more important than what you’re putting in. If you have no particular case for why the one particular detail you decided to focus on is the crux of the whole situation, your forecast shouldn’t be taken as very reliable.
In Super Bowl XXVII, Commentator 3 made the judgment that the Cowboys were the better team, not in one particular way, but overwhelmingly based on their performance in the season as a whole. The outcome of the game seemed to bear out this judgment.
It might seem strange to ascribe the “outside view” to Magnus, who in this case is the ultimate insider. Obviously, no one could possibly know more about his state of mind and his preparation than he could. But the outside view is less about what you know and more about your perspective. In this quote, at least, he’s taking a resolutely big picture view.
For chess, the outside view starts with the players’ ratings. Ratings aren’t perfect, but they’re the best big picture view of performance we have. More specific details of the match shouldn’t be ignored, but they should be used to adjust from the base prediction generated by the ratings, not as blank slate predictions.
The 1972 Fischer-Spassky World Championship is famous for the bizarre circumstances surrounding the match. The players became uneasy avatars for their respective political systems; Fischer threatened not to participate until the last second; some of the games were played in a back room after Fischer complained about conditions in the playing hall. Factors like these can be used to construct an inside view, but from the outside view the most relevant factor going into the match was the ratings: 2785 for Fischer, 2660 for Spassky. This is a huge difference for top level chess. After many twists and turns Fischer won the match 12.5-8.5. One way of seeing this would be to say, a lot of weird stuff happened, but in the end the stronger player dominated the match.
Of course, this would erase a lot of what made the match so fascinating. The big picture stuff is important for forecasting, but it’s usually not very exciting. “Fischer will win because he’s higher rated” doesn’t make for a great story. There’s a tension between description and prediction that often gets glossed over in the media. The commentators for the Super Bowl were ostensibly giving predictions, but ultimately their job wasn’t to be right, it was to generate excitement for the game. For them it’s much worse to be boring than wrong. And you can make a case that chess needs more storylines.
So in the spirit of the holiday season, here are three hot takes:
Magnus will play at least one marathon bullet session during the match. This bullet binge will occur after a dispiriting loss and it will be widely taken to signal his mental collapse, but he’ll defy expectations by winning the next game.
Nepo will cut off his man bun in an attempt to change his fortunes.
We will see the London System on the board for the first time in a World Championship match.
Nate, I have been enjoying reading your posts. I agree with your thesis, and it reminded me how pleasantly surprised I was 20 or so years ago when I learned that chess' ELO rating system had been adapted by several sports stats gurus.
I want to point out that the 1993 Super Bowl was not, in terms of performance, as lopsided as the score suggests. The two most important factors in that game were the injury to Jim Kelly, the Bills' starting quarterback (the Bills were only 7 points behind when he left the game) and 9 (nine!) turnovers by the Bills, neither of which are good measures of football performance. One good (though imperfect) simple metric of football performance is total yardage, and the Cowboys only outgained the Bills by 408 to 362 yards. The Bears victory over the Patriots in the 1986 Super Bowl was a little more lopsided in score (46-10) and a lot more lopsided in yardage (408-123).
I know it's not relevant to your main point, but for the sake of accuracy I have to mention that your paraphrase of commentator #1 must be way off the mark. At the time of Super Bowl XXVII, the Cowboys had not only appeared in 5 previous Super Bowls (V, VI, X, XII, and XII), but they had indeed won 2 of them (VI and XII)