Sector Radar
Sector Radar ranks native US sectors or a bounded set of native US industries before you drill into individual stocks. It uses a bundled deterministic scorer, keeps every active dimension comparable across the scan, and never substitutes ETF returns for the native universe.
Creator
Driven
Created time
—
Last update
—
Version
—
Usage
Built-in skill
How it works
Momentum (60%)
A weighted blend of 63-session (40%), 126-session (30%), and 252-session (30%) compounded proxy returns. Each daily input is the arithmetic mean of constituent percentage changes, so the compounded series is a synthetic constituent-average path — not an investable ETF or official sector-index return.
Acceleration (20%)
The recent 63-session proxy return minus the prior non-overlapping 63 sessions. This separates groups whose relative trend is strengthening from those whose momentum is fading.
Valuation (20%)
Lower positive PE ranks higher within the scanned set. Valuation participates only when every group has a positive PE on one snapshot date no more than seven calendar days behind the shared history date.
Comparable-data guardrails
Equal observations receive equal percentile ranks. A dimension with missing or incompatible data is dropped for the entire scan and the remaining weights are renormalized; missing values are never converted to neutral scores. Fewer than two common return horizons produces an insufficient-data result instead of a partial ranking.
Trend-quality quadrant
Every scored group is classified as strong or weak × accelerating or decelerating. The result supports a qualified allocation interpretation while keeping the evidence date, active and dropped dimensions, warnings, and proxy limitation visible.