Saturday: the only slot where the week is complete and nothing is at stake.

The two conditions

Complete, and consequence-free.

An efficiency review needs two things that rarely coincide. It needs the week to be finished, because a week reviewed mid-flight is a week whose remaining trades will be influenced by the review — which converts a diagnostic pass into a live intervention. And it needs no immediate decision hanging on it, because a review conducted an hour before the session opens will be read for permission rather than for information. Saturday is the only slot satisfying both. The market is closed, every trade has resolved, and the next decision is far enough away that the reading cannot be bent toward it.

The paper bridge

The Trade Capture Document exists because Saturday is late.

There is a cost to reviewing weekly: excursion facts are perishable. MAE and MFE are properties of the price path, and reconstructing them days later from memory or from a chart is unreliable in a way that is difficult to detect afterwards. The Trade Capture Document is the answer — a paper record filled at the time of the trade carrying MAE, MFE, pips risked, duration, fees, swap, and notes, held until the Saturday entry. It exists precisely because CP3 entry is deliberately deferred to the weekend, and it means the deferral costs nothing in fidelity. Broker and TradeZella records serve as the digital cross-check on the same facts.

• Excursion data captured at the trade is evidence; excursion data reconstructed on Saturday is estimation.
• The paper document is what makes weekly batching safe rather than lossy.
• The pips-translation check is the stamp that catches a row where the reconstruction went wrong.

Against reviewing daily

A daily efficiency review would grade noise and change behaviour mid-week.

The obvious alternative is to review continuously, and it fails on both sample and interference. On sample: a day carries too few trades for any efficiency count to mean anything, so a daily read is dominated by whichever one or two trades happened to occur. On interference: a daily review necessarily reaches the operator while the week is still running, and an operator who has just read that their capture efficiency is poor will manage the next trade differently — not because the plan changed but because the reading arrived. That is a live intervention dressed as analysis, and it corrupts both the trade and the following week's data.

Against reviewing monthly

The month is the structural unit; the week is the correctable one.

The opposite error is to batch efficiency review into the monthly cycle alongside the regime classification. That fails for a different reason: an execution problem left running for four weeks has produced four weeks of contaminated data and four weeks of reinforced habit before anyone sees it. Execution is a behavioural layer and behaviour corrects fast, which means it deserves a fast cadence. The month remains the right unit for structural and regime questions, where the underlying signals genuinely need that long to say anything. The two cadences coexist because they are measuring things that move at different speeds.

The key idea

The review's timing determines whether it is a diagnosis or a decision.

The same analysis, performed at a different point in the week, becomes a different act. Read on Saturday it is a diagnosis of a completed period. Read on Tuesday it is an input to a live position; read on Sunday evening it is a search for permission. The content does not change — the slot does, and the slot is what the discipline is protecting.