The week quality grade: counting states rather than averaging numbers.

The rollup

One row per week, built from counts.

Weekly_Rollup aggregates the trade log across the year, and its quality-flag group is where the grade is formed: high-MAE count, elite-MFE count, breakeven trigger rate, plan adherence, and the resulting week quality grade with a review note. Alongside sit the sample and outcome columns, the excursion and efficiency block, friction and duration, and the pips translation. The structure matters — sample first, then execution quality, then friction, then the quality flags — because it is also the reading order. A week with too few trades cannot support a grade, which is why trade count sits at the front rather than being an afterthought.

Counting against averaging

An average is a place for a bad trade to hide.

Consider a week of forty trades in which thirty-four were clean and six consumed almost the entire stop before working. Averaged, the week's MAE reads slightly elevated — a nudge, easily attributed to conditions and easily dismissed. Counted, the week produced six high-MAE trades, which is a specific number of specific rows the operator can open and examine. The information content is enormously different and the underlying data is identical. This is the general property of averaging in a small sample: it converts a locatable problem into a diffuse one, and diffuse problems do not get investigated because there is nothing to investigate.

• A count points at rows. An average points at a feeling about the week.
• Small samples make averages worse, not better — one outlier moves a forty-trade mean substantially.
• The counts are also what make weeks comparable: six high-MAE trades is the same fact in a good week and a bad one.

Adherence in the rollup

Plan adherence sits beside the efficiency counts on purpose.

The weekly grade includes adherence alongside the excursion statistics, which places a behavioural input directly next to the execution ones. The pairing is deliberate and diagnostic: poor efficiency with strong adherence points at the plan, since the operator did what they were told and the results were poor. Poor efficiency with weak adherence points at the operator, and the plan has not actually been tested this week. Strong efficiency with weak adherence is the most dangerous combination in the table, because it is the one that teaches an operator that the rules are optional — a lesson that is reinforced for exactly as long as the improvisation happens to be working.

What the grade cannot do

It grades the week's execution, not the week's edge.

The boundary is the same one that governs the whole lab. A strong week quality grade says opportunity was captured efficiently and the record is clean; it does not say the strategy has an edge, and it certainly does not authorise more risk. The lab reads trade facts after execution and identifies leaks — it cannot grant permission to increase size, ignore a throttle cap, or bypass a gate. If execution quality is improving while the gate state is restrictive, the gate still wins. A run of excellent week grades inside a Floor gate is an encouraging fact about conduct and changes nothing about deployment.

The key idea

Grade the week from what happened to individual trades.

The week is the smallest unit that carries enough trades to say something and stays short enough to act on. Building its grade from counts rather than means keeps the individual trades visible inside the aggregate, which is the only arrangement under which a weekly review can end in opening a specific row rather than in a general impression.