Learning by thinking: what a fifteen-minute post-mortem does that another game does not

Hang around a tournament hall after a round and you'll see it. Two players who spent the last few hours trying to wreck each other are now hunched over the same board, pushing pieces around and arguing about where it went wrong. That's the post-mortem, and it's about as old as chess. Masters do one after every game. Club players mostly don't. The opponent has gone home and the stronger friend isn't picking up the phone.

There's real science behind the ritual, and it says something more specific than “experience is good.” On its own, experience is a pretty weak teacher. The research keeps landing in the same place: playing a lot helps less than you'd expect. What helps is going back over what you played.

Chess psychology really starts with Adriaan de Groot. His study began in the 1930s and came out in English in 1965. He had players from club level up to grandmaster think out loud while picking a move, and what he found still surprises people. The masters didn't calculate deeper or wider than the amateurs. They read the position better, and strong candidate moves showed up for them almost right away.¹

William Chase and Herbert Simon then showed where that comes from. Masters who got a few seconds' look at a position from a real game could set it back up on the board. Novices couldn't. Scramble the pieces into a random position, though, and the masters' edge almost vanished. Their argument was that expertise is a huge store of familiar patterns.² Later work refined those patterns into more flexible templates. The basic conclusion held up anyway: a master is someone who has understood a lot of positions and held on to them.³

So that's the job of a post-mortem. It's the step where a position you played turns into one you actually understood, and only those stick.

Anders Ericsson's well-known work on expert performance argued that experts are made through deliberate practice: effortful, structured work aimed at getting better, with feedback. Just playing a lot didn't qualify.⁴ The strong version of that claim didn't hold up. A meta-analysis found deliberate practice explains roughly a quarter of the variance in performance in games, and less in other fields. Practice is necessary, just not sufficient.⁵

The chess-specific numbers are still worth a look. In two samples of rated players, the hours someone had put into serious study on their own predicted rating better than tournament games did. Better than coaching, too.⁶ And a longitudinal study found the slowest player to reach master level had needed roughly eight times as much practice as the fastest.⁷ Put those side by side. Hours matter. The games themselves matter less than most of us assume, and a lot of that huge gap comes down to how they spend their time. Which leaves the useful question: what should that study actually look like?

The cleanest test of what reflection does isn't from chess at all. Researchers at Harvard Business School ran a field experiment inside a large company's training program. One group of trainees spent the last fifteen minutes of each day writing about what they'd learned. The control group used those fifteen minutes to do more of the training. On the final assessment, the group that reflected scored about 23 percent higher.⁸

These were call-center trainees, and it's a working paper, so we're not going to pretend it says anything about your rating. But it's the same trade-off every improving player runs into, and most of us get it wrong. You just finished a game. You can spend fifteen minutes thinking about it, or fifteen minutes on the next one. In the studies we know of, the thinking came out ahead.

Chess coaches have been saying this for decades. Alexander Kotov's classic manual on thinking at the board spends real time on self-assessment, since you have to know your own tendencies before you can fix them.⁹ And even experts pick better moves when they're made to deliberate instead of going with their first instinct.¹⁰ Strong players benefit from slowing down too.

When nobody's around to ask

What the fifteen minutes look like in Chessalyz

Most club players skip the post-mortem for a boring reason. It needs a stronger player with time to spare, and that person is rarely around at eleven at night after you've just dropped a blitz game. We built Chessalyz to ask that person's questions whenever you want them. Upload the game, or let it sync from Chess.com or Lichess. Two to five minutes later, Matty has found the two or three moments that decided it and asks you about the first one. You write down what you were thinking. Matty reads your answer and coaches from there. Those moments turn into flashcards, the game review shows where you left your opening prep, and your explorer shows what opponents actually play against you.

The questions are modeled on the ones FM Dalton Perrine asks in his own lessons. We never set out to replace him. We just wanted the post-mortem to happen, every game, for the players who would otherwise never get one.