Reading Normal first: the deterioration signals that arrive before the P&L does.

The healthy signature

What a working Normal branch looks like in the MAE/MFE Lab.

A healthy Normal branch has a distinctive execution fingerprint: stable and moderate maximum favorable excursion, high capture efficiency, low giveback, and shallow maximum adverse excursion. The reason is structural rather than fortunate — static exits at fixed R-levels cannot surrender open profit the way a trail can, so capture stays high almost by construction. Which means that when capture efficiency falls on the Normal branch, the exit structure is not the culprit. Something upstream has changed.

The three red flags

High MAE, high MFE with low capture, and the integrity review.

Three signatures are called out by name in the lab manual, and they point in three different directions. High MAE on Normal means trades are being entered into more adverse movement than the branch's structure assumes — an entry-timing or regime-selection problem, not an exit problem. Very high MFE paired with low capture means the trades are running well past 2R after closing, which usually means trend behavior is being labeled Normal. And a rising frequency of branch-integrity reviews is the system telling you the classification layer itself has drifted. That third one is the most serious, because it corrupts every other branch's sample at the same time.

Why the order matters

Normal's state is the denominator for every other branch's read.

Weekly review is not a set of independent branch checks that happen to be performed in sequence — the order is load-bearing. If Normal is healthy and Trend No-Partial is struggling across eleven trades, the honest conclusion is usually 'insufficient sample, continue' rather than 'branch broken, intervene'. If Normal is deteriorating, that same TNP wobble reads completely differently, because a system-wide problem produces branch-level symptoms everywhere and the smallest samples show them least reliably. Reading the fat-tail branches first inverts the diagnostic logic: it lets the noisiest, least-populated branch set the emotional tone for a review of the branch that actually carries the account.

• Normal healthy + trend branch weak → likely branch-specific, likely variance, usually continue.
• Normal weak + trend branch weak → system-level read; branch-level intervention is premature.
• Normal weak + trend branch strong → check for classification drift before celebrating the trend branch.

What deterioration is not

A bad Normal week is not a signal to reach for a more aggressive branch.

The most damaging response to a soft Normal stretch is a rotation toward the trend branches — the reasoning being that if the stable branch isn't producing, the system needs the branches that can produce more per trade. This is exactly backwards, and the gate architecture is built to prevent it. Deteriorating structure reduces variant aggression rather than increasing it; that is the SDE lens doing its job. A soft Normal branch means the system's most reliable evidence source is signalling weakness, which is the worst possible moment to add variance. The correct response is a smaller Normal, not a larger Trend No-Partial.

Sample discipline

Normal earns a shorter read window than any other branch — but not an instant one.

Because Normal accumulates trades faster than the rest of the stack, it reaches decision-useful sample sooner, and the temptation is to treat that as licence to react quickly. It isn't. A soft fortnight on the core branch is still a small sample with real dependence in it — the trades cluster in the same sessions, the same pairs, and often the same regime, so the effective information count is well below the ticket count. What the higher sample rate actually buys is the ability to distinguish a persistent shift from a cold streak within a review cycle or two rather than a quarter. The read is faster than the other branches. It is not immediate.