Review on a Schedule, Not After a Bad Day
Most reviews of a trading approach are not scheduled. They are provoked. Something goes badly, the record gets opened, and an hour is spent looking for the cause. That process feels rigorous and is structurally biased in a way that is hard to see from inside it, because the thing being examined and the reason for examining it are the same event.
Why the Trigger Corrupts the Review
A review that begins after a loss begins with a conclusion already implied. Something went wrong, therefore something is wrong, therefore the task is to find it. The record is then read as a search for a defect rather than as a description of what happened, and a record read that way will always produce a defect, because any sequence of trades contains patterns if you look for them with sufficient motivation.
The same record read on an ordinary morning, with no particular result in mind, produces different conclusions. Not because the data changed but because the question did. That difference in output from identical input is the whole argument for taking the trigger out of the reviewer's hands.
What a Fixed Interval Changes
Choosing the interval before knowing what it will contain means the review after a good month and the review after a bad one are the same procedure. Good stretches get examined too, which matters more than it sounds, because a strategy that made money for reasons unrelated to its logic is a real risk and nobody investigates it voluntarily.
A schedule also gives the urge to intervene somewhere to go. During a difficult stretch the pressure to change something is considerable, and the two usual outcomes are acting on it immediately or suppressing it until it emerges as an unplanned deviation mid session. A scheduled date is a third option: write the observation down, keep following the rule, and take it up when the review comes round.
Choosing the Interval
The interval has to be long enough to accumulate a meaningful number of sessions and short enough that a genuine problem does not run unchecked. For an approach taking roughly one trade per session, weekly is generally too frequent, because a week contains too few observations to say anything and the review becomes a discussion of individual trades.
Monthly or quarterly gives the sample time to mean something. The exact choice matters less than the fact that it was made in advance and is not renegotiated when a stretch turns bad. An interval that quietly shortens during drawdowns has become a result triggered review again, with a calendar attached for appearances.
What the Review Should Ask
Fixed questions, asked the same way each time, so that answers can be compared across reviews. Whether the rules were followed as written, and where they were not. Whether any rule turned out to be ambiguous in practice. Whether the conditions the setup depends on still describe what is being traded. What the distribution of results looks like, rather than the total.
Notably absent from that list is whether the period was profitable. It is worth knowing and it is a poor place to start, because it dominates everything read after it. A review that opens with the result will interpret every subsequent observation through it, and the observations are the part that carries the information.
Emergencies Still Exist
None of this argues for ignoring something genuinely broken until the calendar allows. A rule that cannot be executed, a defect in how a level is calculated, a change in the instrument that makes the setup inapplicable, all of these get dealt with immediately and none of them require a performance argument.
The distinction is between a defect and a result. Defects are structural, they can be described without reference to profit and loss, and fixing one is not a strategy change so much as a correction. Results are the ordinary output of an uncertain process, and they wait for the scheduled date, however much they do not feel like waiting.