Adjust A Strategy Or Leave It Alone

Most strategy changes are made in the wrong week. They get made on a Friday after a run of losses, or on a Sunday night while the last trade is still fresh. The edit feels like discipline. It is usually the opposite: it is the fastest available way to stop feeling uncomfortable.

The question most people type into a search bar is when should I change my trading strategy. That question cannot be answered by looking at the last few trades, because the last few trades are the one piece of evidence your judgement is least able to read honestly. What can be answered is a narrower question: is this a rule failure, or is this ordinary variance? Those two things look identical on an equity curve and are treated completely differently.

This guide sets out an if-then test for telling them apart, and a change log that makes every edit traceable after the fact.

Any strategy with defined rules produces losing runs. They are a property of how outcomes distribute, not a message from the market. A strategy that takes a specific setup in a specific session will have weeks where that setup appears and does nothing, and weeks where it appears three times and fails three times. Nothing has broken. The sample is just short.

Three losses in a row is a sample, not a verdict. The sentence that usually follows a bad week is: this has stopped working. That sentence treats a handful of outcomes as proof, and it arrives at the exact moment your judgement is worst.

The defence is simple and it has to be built in advance. Decide, before the drawdown, what counts as a review sample. Thirty executed trades is a common choice. A full month is another. Whichever you pick, write it down while nothing is wrong, because after a losing run you will find a reason why the sample you agreed to no longer applies.

Until the sample is complete, the strategy is not under review. It is being collected.

Rule failure is structural. It shows up as something you can point at and describe in a sentence, not as a feeling that the market has changed character.

The first form is absence. The conditions the strategy waits for stop occurring. Not failing — occurring. If your setup needs a clean sweep of an overnight high and the market has spent a month grinding sideways without producing one, the strategy has not lost its edge. It has lost its input. That is worth noting, and it is not fixed by loosening the entry.

The second form is ambiguity. Two honest readings of the same chart produce opposite trades. If you can look back at a week of screenshots and genuinely argue both sides of your own rule, the rule is not a rule. That is a drafting problem, and it is the most common thing people misdiagnose as a strategy problem.

The third form is cost. The spread, the slippage and the commission are taking a share of the distance the plan was built around. A strategy designed for a move that no longer clears its own cost is a design problem, not a variance problem.

The fourth form is schedule. The strategy needs you in front of the chart at hours you do not have any more. A plan you cannot execute is not a plan you can evaluate.

Notice what is not on that list: a bad week. A bad week is not a form of rule failure. It is the price of having rules.

Run three questions in order. The order matters, because each one disqualifies the next.

First: did I follow the rules on every trade in the sample? Go through the journal and mark each trade as in-plan or out-of-plan. If a meaningful share of them are out-of-plan, stop. The strategy has not been tested yet. What has been tested is your execution, and the fix is a process fix, not an edit to the rules. Changing a strategy you never actually traded is how people end up on their fifth system with no record of any of them.

Second: if the trades were in-plan, is the sample complete? If you are nine trades into a thirty trade review, the answer is to keep going. Leave it alone. This is the branch people hate, and it is the correct answer more often than any other.

Third: if the trades were in-plan and the sample is complete, can I name the thing that changed? Not sense it — name it. The setup no longer appears. The rule is ambiguous at step four. The average cost has grown relative to the distance I am targeting. If you can write the cause in one sentence, you have grounds for one edit. If you cannot, you have a feeling, and a feeling is not a reason to redraft a process.

An edit made on Tuesday because Monday hurt is not a change to the strategy; it is a change to your mood.

The reason to log changes is not tidiness. It is that strategies drift. Six small edits made over six months, each one defensible on its own, produce a system nobody wrote and nobody tested. Without a log, you cannot tell which edit caused the drift, and you cannot go back.

A change log entry needs five things. The date. The one variable you changed, stated precisely. The evidence that triggered it, in a sentence. What you expect to observe afterwards, described as behaviour rather than outcome — fewer trades before London, say, or no entries taken into a scheduled release. And the review date, with the sample size attached.

Two discipline rules sit around it. One change at a time, because two changes at once tell you nothing about either. And keep the old version intact. If the edit does not do what you expected, you want something to go back to that is not a reconstruction from memory.

If a proposed change cannot be written in one sentence, it is not an edit. It is a different strategy, and it belongs in forward testing on its own, not quietly grafted onto the one you are trading.

This is where having a written reference version earns its place. The strategies material inside the member app lays out a process end to end rather than as a collection of tips. Complete Workflow walks every step from a blank chart to a closed trade in one session. Your Daily SOP sets out the eight-step daily ritual: mark, bias, classify, wait, confirm, execute, manage, journal. Trade Management covers break-even triggers, partial profits, trailing and scaling.

Use it as the comparison document, not as a replacement for your own plan. Put your last sample of trades next to the written steps and check where the two diverge. Most of the time the divergence is at a step that was skipped — the classify step, or the wait, or the journal — and that is execution, which question one of the if-then test already told you to fix before touching anything else.

It also gives your change log a stable reference point. When an edit is logged, you can state it as a deviation from a documented step rather than as a vague adjustment to a system that only exists in your head.

Trading carries risk of loss. Every decision you make is yours, and nothing on this page is financial advice.