Quantitative Signals & Scoring
A quantitative framework that evaluates insider trading signals and market anomalies across multiple time horizons simultaneously, aggregating predictive strength at intraday, short-term, medium-term, and long-term intervals to produce a composite conviction score.
Multi-horizon scoring decomposes a trading signal or insider activity pattern into constituent predictive power across distinct time buckets. Rather than assigning a single scalar score, the methodology computes alpha decay trajectories, signal persistence metrics, and hit rates for each horizon, then combines them through weighted averaging or ensemble techniques. This approach captures both fast-decaying microstructure edges and slower-moving fundamental or behavioral anomalies that insider transactions may trigger.
In insider trading detection and compliance frameworks, multi-horizon scoring is particularly valuable because insider transactions often exhibit different predictive signatures depending on holding period assumptions. A Form 4 filing may signal strong short-swing profit recovery risk (short horizon), moderate sector momentum (medium horizon), and weak long-term fundamental repositioning (long horizon). By weighting these horizons according to market impact decay function and historical edge half-life, quant platforms reduce false positives and better calibrate position sizing and monitoring intensity.