The engine can't teach you. Here is what a question does that +1.3 cannot.

+1.3 is the most honest thing an engine will ever tell you. It is accurate to a tenth of a pawn, it arrives in milliseconds, and it contains no information about you at all. It does not know why you played the move. It has no theory of what you were trying to do. It cannot tell you which of your habits produced it, and it will not notice when you do it again.

This post is about the difference between a grade and feedback, and about why the specific thing a coach does, asking you a question before telling you anything, is not a courtesy. It is the mechanism.

John Hattie and Helen Timperley's review of feedback research is the one most teachers have read, and its central claim is simple: feedback is among the strongest influences on learning, and its power depends entirely on what it is about. Feedback that tells you how you did on the task, how your process went, and how to regulate yourself next time works. Feedback aimed at the person (“good job”) does almost nothing. Effective feedback, they argue, answers three questions: where am I going, how am I going, and where to next.¹

A later meta-analysis found the average effect smaller than Hattie's early syntheses suggested, and hugely variable, for exactly that reason: most feedback is not about the process.²An evaluation answers “how am I going” in the narrowest possible sense and leaves the other two questions blank. It is a score on the outcome of a decision whose reasoning it never saw.

Nate Kornell, Matthew Hays and Robert Bjork ran a study in which people had to guess the answer to a question before being shown it, under conditions where the guess was almost guaranteed to be wrong. The failed attempt still improved later memory for the correct answer compared with simply studying it.³Janet Metcalfe's review of the errors literature reaches the same conclusion from many directions: making an error and then receiving the correction beats being handed the answer, and instruction designed to avoid errors is likely counterproductive.⁴

The mechanism matters for chess. When you commit to a reason, “I took on e3 because trading off that knight looked like the simplest way out,” you have created the thing the correction attaches to. When the coach then shows you that Black recaptures with check and picks up the knight on f7 a move later, the correction has an address. When the engine shows you Qe6 with no commitment on your side, the correction is filed nowhere in particular.

The confident mistake

18. Nxf7 Ne3 19.Bxe3

MovesSummary

You spent a real think on Bxe3. What was your evaluation of the resulting trades, and what other candidate moves were you comparing it against?

You

This seemed to be the simplest reply to prevent the fork on the rook and the queen by exchanging pieces. I get rid of that dangerous attacker, and I could not see an immediate mating sequence that Black is threatening.

After 19...Qxe3+, Black has the forcing check that turns your intended exchange into a material loss: 20.Kd1 Nxf7 21.Rc1 Kh8 22.Re1 Qf4 23.e5. You correctly saw that Bxe3 stops ...Nxd5, but the queen recapture with check comes first. Before any capture, list your opponent’s checks, then their captures, and follow the line until the exchanges have ended.

Try this instead:Qe619.Qe6 was the tactical resource that kept the initiative, attacking the knight on e5 while setting up the exchange on f7.

Your surest blunders are the easiest to fix

Brady Butterfield and Janet Metcalfe expected that errors made with high confidence would be the hardest to correct. They found the opposite. On general-knowledge questions, the wrong answers people were most sure of were the ones most likely to be corrected on a later test, once they had seen the right answer.⁵The finding, called hypercorrection, requires corrective feedback and has mostly been shown with trivia, so we will not stretch it further than that. But it suggests something practical: the move you were certain about, and were wrong about, is the one you will most reliably stop playing, provided somebody makes you state the certainty and then shows you the correction.

That is why a coach asks how sure you were. The engine never will.

Thinking about thinking

18. Nxf7 Ne3 19.Bxe3

MovesSummary

You spent a real think on Bxe3. What was your evaluation of the resulting trades, and what other candidate moves were you comparing it against?

You

This seemed to be the simplest reply to prevent the fork on the rook and the queen by exchanging pieces. I get rid of that dangerous attacker, and I could not see an immediate mating sequence that Black is threatening.

After 19...Qxe3+, Black has the forcing check that turns your intended exchange into a material loss: 20.Kd1 Nxf7 21.Rc1 Kh8 22.Re1 Qf4 23.e5. You correctly saw that Bxe3 stops ...Nxd5, but the queen recapture with check comes first. Before any capture, list your opponent’s checks, then their captures, and follow the line until the exchanges have ended.

Try this instead:Qe619.Qe6 was the tactical resource that kept the initiative, attacking the knight on e5 while setting up the exchange on f7.

Did you notice Qxe3+ was waiting for your opponent after this?

I did not see this replyI saw it but thought I was still okayI saw it but calculated the line wrongNot sure / don’t remember

The skill the engine gives you no practice at

John Flavell gave the field its name in 1979: metacognition, the knowledge of and monitoring of your own thinking.⁶A player with good metacognition knows when they are calculating and when they are guessing, notices when a plan has quietly changed, and can tell the difference between “this move is good” and “I want this move to be good.” None of that is a chess skill exactly. All of it decides games.

You cannot practise monitoring your thinking if nothing ever asks you to report it. A question is a monitoring exercise. It forces you to observe what you did and put it into words, which is the only way to find out it was not what you thought.

Even experts benefit from being made to slow down and deliberate. In a study that used engine evaluations to score moves, both strong and weaker players chose objectively better moves after extra deliberation than on their first instinct, on easy and hard problems alike.⁷

To be clear about what we are not saying: passive review is not worthless. In the classic testing-effect study, students who reread a passage did better than self-testers when quizzed five minutes later; the reversal came a week out.⁸The engine is the most accurate evaluator of a chess position that has ever existed, and Chessalyz runs one underneath everything it does. The point is who speaks first.