Efficiency zones: turning excursion data into a weekly grade.

The compression problem

Precision that never gets read is precision wasted.

The lab computes MAE quality, MFE quality, capture, giveback, adverse utilization, and stop efficiency for every trade — genuinely useful per-trade detail that, at the live rhythm, accumulates faster than any weekly review can absorb row by row. The zone layer solves it the way the gate ladder solves drawdown and the scorecard solves EV: by classifying rather than reporting. Each trade's ratios resolve to an efficiency zone; the zones aggregate into weekly counts; the counts produce a week quality grade. The operator reads the grade first, the counts second, and descends to individual rows only when something warrants it — which is the only reading pattern that survives contact with a busy trading week.

The weekly rollup

What the week's execution actually reports.

The weekly rollup aggregates by week across the year: trade count, wins, losses, win rate, average and net outcome, average MAE and MFE, capture, giveback, fee drag, swap count and fees, duration, pips risked and net pips, high-MAE count, elite-MFE count, breakeven trigger rate, plan adherence, and a week quality grade with a review note. Read in the manual's prescribed order, that's a complete weekly execution picture: sample size first, then whether gross and net outcomes are healthy, then the excursion and efficiency block diagnosing entry precision, opportunity generation, capture, and giveback. The counts matter as much as the averages — a week with three high-MAE trades and one elite-MFE trade has a shape that its averages would completely obscure.

Grading honestly

The grade describes execution, not results.

The week quality grade is an execution verdict, and keeping it separate from performance is what makes it useful. A losing week executed with precise entries, controlled MAE, strong capture, and full adherence should grade well — and the lab's own scenario guidance says exactly that: a red week with high adherence and controlled MAE may be normal variance rather than system failure, checkable against the scorecard, gate state, and benchmark before any rules change. Conversely a profitable week built on trades that nearly stopped out, gave back most of their excursion, and violated the plan should grade poorly, because the grade is answering whether the execution was repeatable, not whether it happened to pay this time.

The escalation path

Grades direct attention; they don't diagnose.

A weak grade's job is to point, not to conclude. From it, the review descends: which block dragged — excursion, capture, or friction? Which counts are anomalous? Which individual trades sit in the worst zones, and do they share a branch, session, ATR state, or variant? That descent is where the diagnosis actually happens, and it usually terminates in one of the layers already built for it — the variant attribution's checkpoint battery, the conditional context slices in the EV lab, or the volatility layer's coefficient audit. The efficiency grade is deliberately shallow so it can be read every week; the depth lives one level down and gets visited only when the grade says to.

• Grade weekly, descend selectively. Reading every trade's zone every week is how the practice gets abandoned.
• Counts before averages when something looks off — outliers hide in means and show in counts.
• Small-sample weeks get a grade with a caveat; the lab's QA rules warn explicitly against overinterpreting thin branch slices.

The key idea

Compression is what makes measurement sustainable.

Every layer of MARS solves the same problem at its own scale: raw data is true and unusable, so it gets classified into states an operator can read at the cadence the data arrives. Drawdown becomes gates; EV becomes green, yellow, red; excursions become zones and grades. The sophistication is preserved underneath and consulted on demand — but what makes the week reviewable is that the top layer is short enough to actually read.