Methodology How the Interest Score works
Not every disclosed trade matters equally. The Interest Score ranks trades by how noteworthy they are — so a first-ever six-figure position stands out from a routine index-fund purchase. It's an attention filter, not a return forecast (see "What the score is" below). This page documents exactly how it's computed. Nothing here is a black box. Current version: v1.1 — every change is logged on the score changelog.
Every input below is knowable at the moment the trade is disclosed. The score deliberately excludes what the price did after the trade. Folding in the outcome would be look-ahead bias — ranking a trade as "interesting" partly because we already know it worked out. That would make the score look impressive and mean nothing. Post-trade price movement is tracked separately, as an outcome (see below), and never influences the ranking.
The 6 components
Each component is scored 0–1, then combined using the weights below (out of 100) into a raw score. That raw score is then percentile-normalized across all disclosed trades, so the final 0–100 Interest Score reads as "more notable than X% of all disclosures" — a score of 90 means the trade is more noteworthy than 90% of the rest. This keeps the ranking exactly as the weights produce it, while making the number itself meaningful and using the full scale.
Relative Conviction — a conditional seventh factor
A $50,000 trade is a rounding error for a member worth $100 million and a 10%-of-everything bet for one worth $500,000. From v1.1, when we can estimate a member's wealth, the score folds in Relative Conviction: the trade's size as a share of that member's own estimated net worth (0–1, saturating once a trade reaches ~10% of net worth). It is fully ex-ante — it uses the wealth disclosed at or before the trade, never a later filing. Net worth is estimated from the annual Financial Disclosures, which report assets and liabilities only in ranges, so it is itself a range. Crucially, a data gap never moves a ranking: for members we have no estimate for, this factor is left out entirely and the score is exactly the 6-component score above — no neutral fudge, no penalty. See the changelog for the full note.
Company & macro context on the signal cards
On the Concentration page, each ranked stock opens into a three-column breakdown: the political anomaly (the factors above), the company fundamentals, and the sector & macro context. The fundamentals (P/E, P/B, EV/EBITDA versus the sector median, revenue growth, debt-to-equity, market cap) are standard multiples computed from SEC-reported financials. The macro column carries the sector's price strength versus the market and the sector's recent congressional money flow (both whole-sector figures, not company-specific), the current fed-funds regime from FRED with a rule-based sector rate-sensitivity, and — only where a sector maps cleanly to a federal department — that department's budget trend from USAspending.gov (a department-level figure, never a company-specific one). This context is shown alongside the ranking to help you judge it; it is not part of the Interest Score. Analyst price targets are deliberately not shown until we have a licensed source for them.
The outcome, kept separate — and what it tells us
We also measure how the price moved after each trade — up to 90 days, direction-aware (a rise after a buy, a drop after a sell), saturating at a 20% favorable move. This is an ex-post outcome: it tells you what happened, but it is not part of the Interest Score. Keeping it separate lets us honestly test the score against outcomes it never saw.
We ran that test on our own data, and we'll tell you the result plainly: a higher Interest Score does not predict a better return. Across thousands of trades, the score's components show essentially no correlation with what the price did afterwards. That's not a bug — it's the honest finding, and it matches what independent research shows: Congress does not reliably beat the market. So the Interest Score is an attention filter, not a return forecast. It answers "which of these thousands of disclosures are worth a look?" — the large, the unusual, the clustered, the ones where the trader oversees the industry — not "which will make money." Anyone who claims a score that predicts congressional-trade returns should be met with skepticism; we show our own calibration evidence so you don't have to take our word for it.
What the score is not
Sources
Trades come from the official US House Clerk and Senate eFD financial disclosures. Sectors are derived from SEC SIC classifications; valuation multiples from SEC-reported financials; legislation from the Congress.gov API. Every trade links to its original filing.