Three layers of review: result, decision, process

1. The result layer: admit that results mislead

A hand is decided by the deal, the draw, opponents’ choices and the information you had. Your decision is only one input. If you use the result as the grade, you will bank luck as skill.

Worse, the result rewrites memory. After a win you remember reading the hand correctly; after a loss you remember “obvious signs” you supposedly ignored. That is the entrance to outcome bias and hindsight.

The first step is not analysis but separation. Write down what happened in the hand without writing where you went wrong. Keep facts and judgements in different columns, or the result will steer everything that follows.

2. The decision layer: only what was visible then

To judge a decision, answer three questions: what could I see, which candidates did I have, and what did each candidate cost? The key phrase is “could see then”, not “know now”.

Example: turn eight, you hold a 2-han hand, an opponent riichi, and you discard an unsafe 5p that deals in. The decision may be wrong or may be reasonable. If that 5p was your only playable tile, or a win would have changed placement, pushing had a reason. If you still had two genbutsu, the decision is the problem.

Writing the candidates is more useful than writing “should have discarded the other tile”. List at least two and state the cost of each: what safety you give up by attacking, what winning chance you give up by defending. Many “obvious mistakes” turn out to be reasonable choices under incomplete information.

The decision layer does not require one correct answer. It requires a reproducible reason. If the same information appeared again, would you choose the same way? If yes, your logic is stable. If no, that is the part to change.

3. The process layer: information, tempo and emotion

Two players of similar level can differ in consistency: one holds up in a tournament, the other collapses late. The difference often lives in process — missed information, tempo stolen by opponents, accumulated emotion.

Information: were you watching discard order, calls and the score? Many errors are not misjudgements but failures to see. In review, ask “did I see it?” before “did I judge it correctly?”.

Tempo: a fast discard after a long think, impatience after consecutive losses, and relaxation after taking the lead are all tempo problems. A fixed action works better than telling yourself to calm down — for example, count safe tiles every time an opponent riichi, then decide.

Emotion: the hand after a deal-in, or the riichi after a losing streak, often carries the previous hand. Marking “what state was I in when I chose this?” is more useful than marking the move simply as wrong.

4. A review template

The template should be simple enough to keep using. Five columns: position, information available, candidates, choice and reason, process note.

Position covers rules, turn, scores and opponent state. Information covers the discards, calls and dora you actually noticed. Candidates: at least two. Choice and reason: what you played and why. Process note: time pressure, emotion, fatigue.

Add a sixth column for the next trigger. Do not write “be careful next time”. Write “when an opponent riichi and I am below 2 han, count safe tiles first”. The more concrete the trigger, the more likely it fires.

Record good hands too. “Why was this well played?” is as important as “what went wrong”; otherwise you only learn to avoid traps and never learn what you are doing right.

5. A 30-minute routine and hindsight bias

First five minutes: pick three hands, chosen for key decisions rather than results. Include a deal-in, a win, and a hand where nothing happened but you hesitated.

Next fifteen minutes: five minutes per hand, filling the template. Facts before judgements, candidates before conclusions. When the thought “I knew that” appears, stop and ask which specific tile you were looking at.

Last ten minutes: turn the common thread into one executable rule and schedule the next practice. For example, “after a riichi, play genbutsu first”, then find ten similar positions in the discard trainer to test it.

The antidote to hindsight is simple: write your judgement before you learn the result. Pause the replay, write “I would discard this and here is why”, then continue. That one habit keeps “I would have known” outside the review.