Trade Review: Grade the Decision, Not Result
A trade review is a structured reading of past trades against the reasons recorded for taking them. It only produces information if the reasoning was written before the outcome was known, because memory reconstructs the reason around whether the trade worked.
How it works
A review reads a record. Not a chart, not a memory — a written record of what was done and why, made at the time.
Reasons recorded afterwards are reconstructions. Once the outcome is known, memory assembles an explanation consistent with it, and the explanation feels like a recollection. That happens to everybody and the only defence is writing the reason down first.
One field does most of the work: what you expect to happen, in a sentence, before entry. Reviewing against that turns a list of outcomes into a record of predictions, which is the only thing that can be graded.
Decision and outcome are separate
A trade has two properties: whether it followed the process, and whether it made money. They are independent, and grading on the second one teaches the wrong lesson roughly half the time.
A rule-breaking trade that made money is the most dangerous entry in a journal. It rewards the exact behaviour a process exists to prevent, and it will be repeated. Grading it as a process failure — while recording the profit honestly — is what stops that.
Two columns, then: process grade and result. Reviewing the four combinations separately is more informative than any single measure of performance.
The half of the sample nobody has
Setups that qualified and were skipped are absent from every broker statement. So is every limit order that did not fill. Which means the reviewed sample is systematically filtered — it contains the trades you were comfortable enough to take.
Logging skipped setups is tedious for a fortnight and then diagnostic. A pattern of skipping trades in one direction, or after a loss, is a finding you cannot reach any other way.
Sample size is the constraint everybody underestimates. Twenty trades cannot distinguish a method with an edge from one without; two hundred begins to. That is why abandoning a method after a bad month is usually a decision made on no evidence.
In practice
Total costs paid over the period belong in every review, as a number and as a share of gross result. It is available from the broker, it is never in a trading journal, and for an active trader it is frequently the largest single line.
Planned stop against actual fill is the second missing number. The difference is slippage, it is measurable, and most traders have no idea what theirs is — which means their risk model is an estimate they have never checked.
Tagging conditions is what makes a review sortable. Session, volume relative to average, whether news was scheduled. Sorting two hundred trades by tag frequently produces a finding that no amount of reading them chronologically would.
A single gap can account for most of a period’s result, which is why reviewing the median trade alongside the total is worth doing — one outcome should not describe a quarter.
A review needs a fixed cadence and a fixed set of questions, or it becomes a mood. Weekly for compliance — how many rules were broken, and which — and quarterly for anything about the method, because a quarter is the shortest period likely to contain enough trades to say anything.
And the questions should be written down once and reused. Which conditions produced the best process grades. Whether costs are rising as a share of gross. Whether slippage has widened. Whether skipped setups cluster anywhere. Five fixed questions asked every quarter produce a comparable series, and a comparable series is the only thing that shows whether anything is actually improving.
What a trade review is not
It is not a performance report. Result is one of the columns, not the purpose.
It is not useful without pre-recorded reasoning. Otherwise it reviews reconstructions.
It is not a monthly event only. Compliance is recorded per trade; analysis is periodic.
And it is not complete without the trades you did not take.
When it fails
In a range every review reaches the same conclusion, which produces pressure to find something new to change. Recognising the regime rather than the method is the harder and usually correct reading.
The second failure is grading on outcome. It rewards rule-breaking that happened to work.
A third is reviewing too small a sample. Twenty trades supports no conclusion, and conclusions get drawn anyway.
A fourth is omitting costs. For an active method they are frequently the largest line and the least examined.
And a fifth is reviewing only losses. The trades that worked for the wrong reasons are where the repeatable mistakes are hiding, and nobody looks at a winner and asks what went wrong.
The original data
On this site’s shared 576-bar history the round-trip cost is 0.0098 price units — 2% of the median bar
range of 0.493 and 45% of the smallest bar of 0.022 — and it exceeds 10% of a bar’s range on 15 of the 576
bars. The figures are in research/series-measurements.json, produced by site/measure_series.py.
The cost line is the one to add to your next review, because it is the one you can act on immediately. Total commissions and spread paid, divided by gross result, for the last hundred trades. If costs are consuming a large share of the gross, the fix is fewer trades rather than better ones — and that is a change you can make this week, unlike an improvement in judgement, which takes a sample size most traders never reach.
Related
Trading journal is the record a review reads. Trading rules is what compliance is measured against. And trading plan is the document a review is testing.
The change that made reviews useful was one column: what I expected to happen, written before I entered. Without it I was reading a list of outcomes and inventing reasons, and the reasons were always shaped by whether the trade had worked.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.