Study Your Best Trades, Not Your Worst (But Not the Way You Think)

Every trader reviews his losses. It feels like the responsible thing to do, it feels like work, and it is very nearly useless.

Brett Steenbarger, who came to trading from brief therapy rather than from the market, brought with him an idea that sounds soft and is not. Stop investigating the problem. Investigate the exceptions, the times the problem did not occur, because the exceptions contain the method and the problem contains only the problem.

The idea has since been flattened into a slogan, study your winners and not your losers, and the slogan is not quite right. Here is the arithmetic, and then the corrected version, which is better than either.

Most of your losses are innocent

Assume, generously, that one trade in five contains a genuine rule break. Trades executed cleanly win forty-five percent of the time. Trades with a rule break win twenty-five percent, because breaking a rule is bad but not fatal.

Your overall win rate comes out at forty-one percent. Now you sit down on Friday and open a losing trade. What is the chance it contains an error?

A loss raises the probability of an error from twenty percent to twenty-five. That is the entire diagnostic power of the thing you have been doing every Friday for years.

This is how traders acquire rules that are not rules. The stop was “too tight” because this one got stopped and reversed. The entry was “too early” because it went against you first. Neither observation would have survived a winning outcome, and neither is a property of the decision.

The slogan does not survive its own arithmetic either

So study the winners instead. Reasonable. Let us check.

Assume exceptional wins arrive fifteen percent of the time from clean execution and five percent of the time from broken execution. Now open your biggest winner of the month. What is the chance it was executed cleanly?

A lift of 1.15. Less informative than the loss.

Which means the popular advice, review your winners instead of your losers, improves nothing. It swaps one weakly diagnostic filter for a slightly weaker one. Both are filters on outcome, and outcome is dominated by variance.

The corrected version

Steenbarger’s actual claim is not about winning trades. It is about exceptions to the problem. And the problem is a behaviour, not a loss.

So the exception is not your best result. It is your best execution: the session where you were tired and followed the plan anyway, the day the market gapped and you sized correctly, the trade you did not take.

Those sessions have a property no outcome has. They are repeatable, because they were caused by you.

This is why the A/B/C session log comes first, and why it must be written before you look at the numbers. Without it you have no way to select an exception, and so you fall back on the only filter available, which is the P&L, which is noise.

The review that works

Why this feels wrong

Because reviewing losses is a form of penance, and penance feels productive.

An hour spent picking over a losing trade delivers something the correct review does not: the sense of having paid for the loss. And it delivers a story with a cause, which relieves the intolerable possibility that you did everything right and lost anyway, which is what happens fifty-nine percent of the time by construction.

The solution-focused review offers none of this comfort. It asks you to look at a session that went well, which feels indulgent, and to find in it something so ordinary that you had not noticed it. There is no absolution in it. There is only method.

What it looks like when it works

The finding is never dramatic. It is always something like: on the A-sessions I had written the plan the night before. Or: on the A-sessions I did not open the platform until the first thirty minutes were gone. Or: on the A-sessions I had already decided which two setups I would take.

These are not insights. They are procedures, and procedures are the only thing that transfers from one week to the next.

The trader who reviews losses collects rules. The trader who reviews executions collects procedures. Rules are things you must remember while tilted. Procedures are things you did before the market opened, when remembering was easy.