The six-table pipeline: one framework, learned once, applied to every metric.

The chain

A signal-processing pipeline wearing spreadsheet clothes.

The stack is a classic processing chain, and each link answers one question the previous link can't. Table 1, Structural Truth: what is this metric right now, and what is its lifetime anchor? Table 2, Rolling Condition: what does it look like across the fast (4W), medium (6W), and structural (12W) lenses? Table 3, Stability: can that behavior be trusted, or is volatility expanding? Table 4, Drift: are the fast lenses pulling away from the slow ones? Table 5, Z-Score: is that drift meaningful relative to the metric's own volatility? Table 6, Interpretation: what does the whole picture actually mean for this branch? Level, smoothing, volatility, first derivative, normalization, classification — the same architecture signal engineers have used for decades, applied to trading structure.

Why identical everywhere

Uniformity is what makes five metrics learnable.

The pipeline could have been tailored per metric — drawdown genuinely behaves differently from expectancy — and the guide deliberately refuses. The table sequence is intentionally consistent across EV, DD, RAER, RAPF, and ACCEL so the operator can read every metric the same way. The payoff compounds: an operator fluent in the EV pipeline is instantly fluent in the acceleration pipeline, the z-band thresholds mean the same thing everywhere, and cross-metric reads become possible because the outputs are structurally comparable. What varies per metric isn't the framework but the semantics inside it — most importantly the sign logic, since for EV, RAER, RAPF and ACCEL positive drift means improvement, while for drawdown the interpretation inverts because worse drawdown is bad.

The stability layer

Table 3 is a trust gate, not a direction read.

The stability table is the one operators most often misread, because it looks like it should say something about performance and deliberately doesn't. It measures noise: σ4 for the fast instability view, σ6 as the medium view and early expansion detector, σ12 as the structural anchor. Its headline output is the σ6/σ12 expansion ratio, and the bands are explicit — below 1.00 is compression and stabilizing, 1.00 to 1.10 is normal, 1.10 to 1.30 means instability is building, and above 1.30 is elevated instability. The ratio's job is to tell you how much weight the rest of the pipeline's readings deserve: a strong drift signal in a metric whose volatility is expanding is a weaker claim than the same drift in a settled series. Trust before direction.

The maturity rule

Blanks, not zeros — the discipline that keeps the pipeline honest.

The guide's first-listed failure mode is deceptively small: using zeros instead of blanks for immature windows, which creates false stability and corrupts rolling and drift logic. A 12-week window at week 8 doesn't exist yet, and filling it with a zero doesn't produce a conservative estimate — it produces a fabricated data point that drags averages, deflates sigmas, and manufactures drift signals from nothing. The rule is blunt: outputs stay blank until the full window exists, so σ12 should not appear before week 12. It's the same statistical-honesty gate as the EV lab's sample minimum and the volatility matrix's status ladder — the pipeline refuses to compute what it hasn't yet earned the right to say.

• Immature windows blank, always. A partial window's average is not a conservative version of the real one.
• σ12 is the z-score denominator — using σ6 there destabilizes the entire interpretation layer.
• Drawdown's cumulative is a running worst, not an average: DD is path-dependent, and average DD isn't max damage.

The key idea

One pipeline is what turns five metrics into a language.

The SDE's real achievement isn't any individual diagnostic — it's that expectancy, capital pressure, risk efficiency, profit quality, and compounding velocity all get described in the same grammar. Value against anchor, three horizons, stability ratio, drift set, normalized strength, categorical meaning. Once that sentence is fluent, the structural layer stops being five workbooks of statistics and becomes one readable account of how the machine is actually behaving.