C+

Market X-Ray · Toolbox 7

Market-Structure Anomaly Score

A single weak ratio is a normal alert. This asks the higher-level question: is today's whole combination of internals — participation, concentration, credit, size and factor/beta — historically unusual? It scores the standardised multivariate distance of today's configuration against the engine's own ~3-year history. This is the 63-day structural lens — distinct from the short-term cracks on the Regime page.

Updated daily · data as of 2026-06-09

9 th pct
normal No structural anomaly Bullish Lifecycle: Resolved · easing — 9th now vs 87th 5d ago

Today's market structure reads broadly normal (9th percentile of abnormality across 5 dimensions) — no notable multivariate anomaly right now. Internals are roughly where the headline would suggest.

What drives it — five dimensions

Each internal dimension's recent move, standardised against its own history (z). The anomaly is the multivariate distance — how far today's whole vector sits from the norm, not any single dimension.

Breadth z -0.32

near its norm

Concentration z +0.24

near its norm

Credit z -0.09

near its norm

Small-cap z +0.98

small caps participating

High-beta z -0.17

near its norm

Statistical context

Abnormality9th pct
For this regime9th pct
Distance (RMS z)0.48
History719bars

How to read it

Five internal dimensions — breadth (RSP/SPY), concentration (ex-Mag-7/SPY), credit (HYG/LQD), small-cap (IWM/SPY) and factor/beta (SPHB/SPLV) — are each measured as a recent ratio move and standardised against ~3 years of their own history. The anomaly is the RMS z across all five: how far today's combined configuration sits from its norm, expressed as a percentile. A high score means the combination is unusual even when no single ratio looks alarming. Direction flags whether the abnormality is a firm-index-over-weak-internals warning or an improving-internals recovery; lifecycle tracks whether it is new, worsening, persisting or resolving. Thresholds are uncalibrated pending backtest, and the distance treats the dimensions as independent (Euclidean, not full Mahalanobis), so correlated dimensions count more than once — directional, not exact.

For information and discussion only — a reading of market internals, not investment advice. Feeds the Daily Read; pairs with Regime Baselines and the Cross-Toolbox Alert Stack.