Tier usage as a diagnostic.

Layer 01 — Earned, not available

T5 and above is acceleration that was earned.

The benchmark's own interpretation sheet states the position directly: the dynamic ladder is preferred because it preserves defensive contraction while permitting controlled acceleration, the upper-middle tiers are earned rather than default, and the top of the ladder is exceptional. A live system reaching the higher tiers substantially more often than the model does is over-aggressive unless the frequency is justified by structural confirmation, and the justification has to exist before the trades, not after the results.

Layer 02 — Per-tier expectancy

High tiers must out-earn their risk, per trade, on record.

The sharpest question in the set is whether trades placed in the upper tiers actually produced better per-trade R than the middle of the ladder. Elevated deployment that earns ordinary outcomes is uncompensated risk wearing a confidence costume, and the account total will never reveal it, because a larger position on an average trade produces a larger number without producing a better one. Only the per-tier breakdown separates the two.

• Riding higher tiers too aggressively stays visible here even while headline returns look good.
• Bigger swings without proportionally better outcomes is the definition of the ladder being climbed for nothing.
• Tier-conditional R is the row that decides whether acceleration was earned or merely taken.

Layer 03 — Under-deployment counts

Authorised tiers left idle is modelled expectancy left on the table.

The diagnostic runs in both directions, which is unusual for a risk instrument and deliberate here. Persistently deploying below what the gate authorises is not free caution. It converts a modelled edge into a smaller realised one while carrying the same operational cost, and over a full horizon that gap compounds in the same way the losses it was avoiding would have. Timidity is a finding, and it gets written down as one.

Layer 04 — Compression, verified

Is the ladder actually contracting when the gate says so?

The last question in the set is whether tier compression during defensive gates is genuinely happening or whether override use is quietly flattening the ladder. This is the failure that looks like nothing from the outside: gate state moves correctly, the panel displays the correct cap, and deployment never actually falls because each individual override felt reasonable at the time. The override stress table exists to make that pattern countable.