The hybrid accelerator: one trade managed as two different trades.

The split

Half the position stops being a trend trade at 1R.

Trend Partial's Standard expression banks 50% at 1R and manages the remainder with ATR-trail continuation logic that activates at 1.6R. The moment that partial fills, the trade has become two things. The banked half has already resolved — it earned a fixed, deterministic 1R and its contribution to the week is settled. The remaining half is now a pure trend instrument with no floor beyond the break-even stop and no ceiling at all. They share a ticket and a symbol, and they share nothing else: different payoff shapes, different variance, different diagnostic questions.

Why the branch exists

The hybrid is not a compromise between two better options.

It is tempting to read Trend Partial as a hedge — half a Normal trade and half a TNP trade for operators who can't decide. The design intent is different. The branch exists because most valid trend setups are not clean enough to justify full exposure but are too good to close at a static 2R. Trend Partial monetizes the part of the move that is reliably available while keeping optionality on the part that isn't. Its 33.75% default weight — second only to Normal, and three times Trend No-Partial — reflects a judgment that this middle condition is the most common trend condition, not that the branch is a fallback.

The coefficient dependency

This is the branch where coefficient selection matters most.

Normal has no trail, so it has no coefficient exposure. Trend No-Partial has a trail but no banked portion, so its coefficient errors show up as clean wins or clean losses. Trend Partial is the branch where a coefficient error is genuinely ambiguous, because the banked half keeps the trade's total R positive even when the runner is being managed badly. A week of Trend Partial trades can post acceptable aggregate R while every single runner is being strangled or surrendered — the cushion hides the defect. This is precisely why the trail coefficient routes through the Volatility Distance Matrix rather than being eyeballed from raw ATR, and why every trail decision is logged so a bad month traces to specific coefficient choices instead of vibes.

• Coefficients too tight → a valid trend is stopped in noise; MFE stays high while capture collapses.
• Coefficients too loose → a mature runner gives back its open profit; giveback rises with MFE unchanged.
• Capture efficiency and giveback are read as a pair — one without the other doesn't identify which error you made.

The friction note

Longer holds make this the first branch where cost quietly compounds.

Normal's holds are short and swap-free by construction. Trend Partial's runner can hold across sessions and across rollover, which makes it the first branch in the stack where fee and swap drag accumulate into something material. The lab tracks fee R drag per branch for exactly this reason. A runner that captures 2.4R gross and 2.05R net is a different instrument from the one the expectancy model assumed, and the gap widens with hold duration. Trend Partial's economics are net economics or they are fiction.

How to diagnose it

Split the report card, because the branch is already split.

The practical consequence of the two-instrument structure is that aggregate branch metrics are the least informative view of Trend Partial available. Blended capture efficiency across banked and runner portions is a number with no clean interpretation, because the banked half captures perfectly by construction and drags the average toward health regardless of what the runner did. The useful reads separate them: partial fill rate as the entry-quality question, and runner capture against MFE as the coefficient question. The MAE/MFE Lab keeps these distinguishable, and an operator reading only the branch-level roll-up will systematically miss the half of the branch where the decisions actually are.