Capture and giveback: grading the exit against what the trade offered.

The invisible number

P&L grades the outcome; MFE grades the opportunity.

A trade that closes at +0.4R looks identical in the journal's outcome column whether it never got further than +0.5R or spent an afternoon at +1.8R before bleeding back. Those are radically different trades: the first was managed about as well as it offered, the second surrendered three-quarters of its opportunity. MFE is the recording of what was actually available — the highest unrealized profit the position reached — and it's captured per trade precisely so the outcome can be judged against it. Without MFE, exit quality is a feeling. With it, exit quality is a ratio, computable per trade, per branch, per variant, per week.

The two metrics

Capture efficiency and giveback — one ratio, one remainder.

The lab computes the pair exactly as defined: capture efficiency is outcome R divided by MFE R for positive outcomes — the fraction of available favorable movement actually collected. Giveback is MFE R minus outcome R, floored at zero — the open profit returned before the exit, expressed in R. They answer complementary questions: capture asks how good the harvest was proportionally; giveback asks how much was left in the field absolutely. The distinction matters at the extremes — a small trade with 90% capture and a monster with 60% capture may carry identical giveback in R terms, and which one deserves attention depends on which question the review is asking. The weekly rollup carries both, alongside their supporting cast: adverse utilization (how much of the stop was consumed), the MFE/MAE ratio (reward opportunity per unit of pain), and fee drag netting the whole picture.

Reading it by structure

Capture is graded against the branch's design, not against perfection.

A naive reading treats 100% capture as the goal, which would make every trailing structure look broken — a trail is a deliberate giveback machine, paying a retracement toll for the chance at trend-length capture. The honest read grades each branch and variant against its own architecture. Normal's static 2R target should show tight capture on completed trades — its opportunity is capped by design. The trend branches should show structurally higher giveback with structurally higher MFE — that's the trade they exist to make, and the question is whether the tail capture pays for the retracement toll over the sample. The variant axes shift the expectation again: Time-Conservative's early rungs buy capture at the cost of tail; Time-Aggressive accepts worse interim giveback hunting larger MFE. The lab's per-branch and per-variant slices exist because 'good capture' has no meaning until you know what structure was being executed.

• Rising giveback within one branch against its own history is the actionable signal — cross-branch comparisons mostly measure design differences.
• Chronic near-perfect capture on a trend branch can itself be a finding: exits so early the trail never gets to work.
• Capture and giveback feed the BE-rule and TP-discipline cost reads in the EV lab — the same leak, priced in EV terms.

The compounding stake

Exit leakage is a tax on every winning trade forever.

The reason this pair earns its own lab is arithmetic scale. A system leaking a consistent fraction of its winners' MFE is paying that fraction on every positive trade, indefinitely — a drag that behaves exactly like fees but rarely gets audited like them. Reclaiming even part of it is among the cheapest expectancy improvements available, because it requires no new edge, no new setups, and no additional risk: the opportunities were already generated and already paid for in MAE and time. The trades happened; the money was on the table; the only variable is the harvest. That's why capture efficiency sits on the dashboard's KPI row rather than in a diagnostic appendix — it's the ongoing bill for management quality, and the lab renders it weekly whether or not anyone wants to see it.

The key idea

The market grades your entries; only you can grade your exits.

Entry quality shows up on its own — hit rates, MAE, win rate all report it unprompted. Exit quality hides inside winners, where nothing complains: the trade closed green, the journal shows a profit, and the surrendered R appears nowhere unless MFE was recorded and the ratio computed. Capture and giveback are the system's answer — a per-trade, per-structure, per-week accounting of the gap between offered and kept. It's the difference between knowing you won and knowing how much of the win you collected.