Methodology

The engine re-reads the public exhaust of federal money every two hours — awards, subawards, solicitations, budget lines, filings, FDA decisions, the Federal Register, patents, sanctions lists — plus market tape (options, sentiment) to detect what's already believed. Each event that maps to a public ticker becomes a candidate signal.

The score, 0–100, weighs three families of evidence: materiality (award size against market cap and revenue), history (what similar awards did to similar stocks, from our own outcome database), and regime (the macro conditions the signal lands in, which historically mute or amplify each class). We don't publish the exact weighting — that's the moat — but every input is shown on the signal's detail page, and Terminal users see each component's contribution to the final number.

Timing is scored separately from strength. A large award read late gets a high score and an honest "likely priced in" flag — conflating the two is how tools trick users into chasing stale prints.

When a signal fires, we record the stock's price and its sector ETF's price. Sixty trading days later, the difference in returns is the signal's alpha — sector-relative, so a rising market doesn't flatter the model and a falling one doesn't punish it.

• Medians, not means. A few microcap moonshots can make a mean look brilliant; the median is the outcome a typical signal actually delivered.
• Fixed buckets, assigned at signal time. A signal scored 68 stays in the 60–71 bucket forever — no re-labeling after the outcome is known.
• Losing buckets published. The track record shows every bucket, including the ones that underperform. A score only means something if low scores predict bad outcomes as reliably as high scores predict good ones.
• Outcomes accrue at the horizon. Recent signals don't count until their 60-day window closes — there is no partial credit for a good first week.
• The loop feeds back. Realized outcomes flow back into the history component, so the model is continuously re-fit against what actually happened, not what should have.