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.
更新于—
内置技能
Driven
内置技能
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.
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.
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.
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.
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.