Take one turn apart carefully.

Choose a useful moment

Start with a recent match and identify the first turn where you remember pausing, changing your mind, or acting automatically. Do not begin with the most dramatic final turn unless that is where uncertainty first appeared. An earlier choice often shaped the later crisis, and it usually has more plausible alternatives to study.

A useful moment contains a decision you could describe without replaying the entire match. You might have chosen between developing pressure and preserving flexibility, between clearing part of a board and accepting risk, or between spending a limited resource and saving it for an expected threat. The exact cards matter less than the structure of the choice.

If no turn stands out, choose the mulligan. Opening choices are short, information is bounded, and the assumptions are easy to write down. The goal of this first session is not to diagnose your whole game. It is to produce one honest field note that you could test again.

Reconstruct facts before explanations

List what was visible at the start of the decision: your available resources, health pressure, board, hand size, known cards, revealed opponent information, and any turn limit that mattered. Mark uncertain memories as uncertain. A blank is safer than a detail invented to make the later story coherent.

Now separate facts from interpretations. Opponent has four cards is a fact. Opponent is holding removal is a hypothesis. My life total is low is a fact. I must clear everything is a proposed consequence. This separation prevents a feeling from entering the record disguised as information.

Hide or ignore what happened after the decision while you compare lines. Later information can test a plan, but it cannot become something you supposedly knew. If you cannot reconstruct enough of the position to compare options fairly, discard this moment and choose another. That is good evidence discipline, not a failed review.

• Visible: board, hand, resources, health, turn, and confirmed effects.
• Known: information revealed earlier and still relevant.
• Unknown: opposing resources or draws that could change the line.
• Assumed: your current model of the opponent's plan.
• Constraint: the one thing your line must accomplish this turn.

Compare two complete lines

Write the line you played, then write the best alternative you can find. Give the alternative the same effort. Describe the full order of actions, the resources left afterward, the pressure created or absorbed, and the next-turn options each line preserves. Comparing one complete line with a vague maybe is not a real test.

For each line, name the favorable response you hope to see and the credible response that punishes you most. Do not search for every possible answer. The useful question is whether the line remains acceptable against the range you reasonably expected. A line with a higher ceiling may still be worse if its common failure leaves no recovery.

Then choose as if the result were hidden. Finish this sentence: I prefer this line because it best protects or advances ____, while accepting ____. If you cannot fill both blanks, your reason is probably a slogan such as tempo or value rather than an explanation tied to the position.

Keep one lesson that can be tested

Reveal the later turns and ask what they teach about your reason. If an unlikely answer appeared, do not automatically condemn the line. Ask whether you assigned that answer a sensible weight. If the opponent lacked an expected response, do not automatically praise the line. Ask whether your plan would still be sound when that response exists.

Write one conditional lesson, not a universal command. For example: when my only flexible answer also protects the following turn, I will compare the cost of taking damage now with the cost of having no answer later. This sentence tells your future self what to notice and what comparison to make. Never spend removal early is easier to remember but far less accurate.

End by scheduling the lesson's next appearance. During the next three matches, pause when the same structure appears and state your reason before acting. You are not trying to force the old answer onto a new position. You are testing whether the decision trace helps you notice the right variables sooner.