Research

How we decide what signals to include — and what to leave out. Every study below is reproducible from public SEC data, with full methodology and sample sizes in plain view.

April 2026 · Signal test

Do insider buys from newly-appointed executives outperform?

Twelve months of Form 4 purchases cross-referenced against 8-K Item 5.02 appointments. Three cohorts, two variants, forward returns vs. SPY at 30 / 60 / 90 days.

SEC Form 4 8-K Item 5.02 Ncohort 1 = 92 · Ncohort 2 = 1,197 Result: no edge (p=0.82)
Read the test →
April 2026 · Non-finding

Why congressional trades don’t overlap with insider buys

We wanted to test whether insider buys at tickers where Congress has also been buying outperform. A pre-registered sample-size check killed the test before we ran it — the two universes barely touch. A short note on why.

SEC Form 4 House PTRs 105 F4 tickers · 158 congress tickers · 1 overlap Result: abandoned at gate
Read the note →
More research coming
Next up: repeat-buyer reliability, cluster buying as a standalone signal, and 52-week-low timing.

How this works

Each research post here tests a specific hypothesis about insider buying — usually one that’s either a candidate for the daily scoring model or a claim we’ve seen repeated without evidence. We publish both the positive and negative results. Methodology and sample sizes are front and center; anything ambiguous is flagged, not hidden.

Get the signals daily

Free. No paywall. Just public SEC data, scored and explained.

Subscribe free ← Back to homepage