Research

Insurers and financials did best, telecom worst, and biotech beat its peers but still lost money. Plus how our alerts compare over the same months (n=27, promising, not proven).

27 alerts with 90 days of data, tested against the S&P 500, the Russell 2000, similar stocks and plain insider buys. A clear edge over the indexes, a thinner one over the rest, and what is not proven yet.

A 2021-2026 backtest: the typical lone insider buy does not beat small caps after costs, a track record does not predict the next trade, and filtering matters more than copying one person.

Every time a corporate officer, director, or major shareholder buys or sells their own company’s stock, they have 2 business days to tell the SEC. Here’s what a Form 4 actually is, what it discloses, and how to read one.

Three executives buying the same stock on the same day sounds like conviction — until you notice they all paid the identical price. Here’s the signature that separates a real cluster from a company quietly selling shares to its own insiders, and how we score it.

Why corporate insiders can't buy or sell stock during certain weeks every quarter — and why that means fewer ClusterSignal alerts around earnings season, not a broken pipeline.

We tracked 26 real BUY alerts with complete 30-day data (updated October 2026): a 73% win rate, and a holding-period pattern that separates winners from losers within the first week.