MAE/MFE Lab — The Repair Engine
The Repair Engine. One concrete fix, not a resolution.
Read this first
Driver diagnostics
01Ranked against targets, flagged as DRAG or WATCH.
Seven execution drivers get measured against their branch-aware targets and flagged when they're costing expectancy:
- MAE pressure and adverse utilization — is the stop being consumed too deeply?
- MFE opportunity — are trades generating enough favorable excursion to pay?
- Capture and giveback — how much of the offered move survives to the close?
- Fee drag and adherence — is friction or discipline the leak?
The What-If engine
02Which single repair pays most?
Instead of fixing everything at once, the What-If Scenario Engine re-computes net EV under single-variable improvements — capture raised to target, giveback halved, fee drag normalized — and ranks the outcomes. The review ends with one named repair and its expected EV payoff.
Discipline boundary
03Repair the process, not the sample.
Sample size gates every read — one or two trades in a timeframe bucket prove nothing. And capture can't be optimized so hard that trend branches get choked: capture and giveback trade off against continuation room by design.
One-change discipline
04The engine issues a single fix per cycle, deliberately.
When efficiency grades slip, the repair engine outputs one concrete adjustment — a stop increment, a partial level, a BE timing change — never a bundle. Single changes keep attribution clean: if the next window improves, the fix earned it; if not, the fix is reverted. Bundled repairs produce improvement nobody can explain and regressions nobody can unwind.
How MARS uses this
Every proposed rule change - stop policy, branch weights, fee model - is cloned into a scenario profile and resampled against the same evidence. The grid renders the trade-offs, and only profiles whose edge survives without deepening the P10 tail earn candidate status for live promotion.
BASELINE
LIVEcurrent governing profile
TIGHTER STOPS
REJECTEDEV cost exceeds drawdown saving
TNP WEIGHT +10
SANDBOXedge up, adverse tail deepens
FEE MODEL B
CANDIDATEfriction saving survives resampling
How it benefits you
System changes stop being vibes-based. The tempting tweak that costs 0.07R of expectancy for a modest drawdown saving gets rejected by arithmetic before it silently taxes six months of trading - and promising candidates carry their evidence with them into review.
Four scenario profiles judged side by side: expectancy, adverse-tail cost, and a verdict. Changes graduate through this grid or not at all.
Further illustration
The MARS operating loop: execute, capture, score, gate, throttle, deploy - every cycle feeding evidence into the next.
Connected inside MARS
This module doesn't work alone.
Go deeper
Operator briefs on this territory.
Deep dive — 01
Every candidate repair has a price in R, and the engine's job is to print it.
Better entries or better exits? The engine answers in R, and the answer is frequently not the loud one.
Read the full brief →
Deep dive — 02
One change per cycle is not caution. It is the only way the next window means anything.
Ship two fixes, get one uninterpretable result. The rule is about evidence, not caution.
Read the full brief →
Deep dive — 03
Driving capture toward one hundred per cent would destroy the branches it was meant to protect.
Perfect capture means exiting at every pause. The metric goes up; the fat tail disappears.
Read the full brief →
Every module ships in the complete MARS package.
One price. Eleven workbooks, three TradingView indicators, and the full manual library — $497.