Methodology
The score answers one question: would copying this trade have been profitable, and is the move still alive?— not “is this stock popular in Congress?” The model ranks; you decide. Every number the score uses is shown in the tables, and /performance records how past picks actually went.
The 0–100 score
A weighted blend of five components:
| Component | Weight | What it measures | Why |
|---|---|---|---|
| Return since disclosure | 33% | Average % gain from each buyer's public entry price (their disclosure date) to now. | The literal “was copying this profitable?” answer — weighted highest on purpose. |
| Distinct buyers | 22% | How many different politicians bought the name (capped at 6). | Several independent buyers signal more conviction than one big trade. |
| Momentum | 18% | Is the price still above its 50-day moving average? | Stops us copying a name that already gained and has since rolled over. |
| Recency | 15% | How fresh the latest disclosure is (45-day half-life). | Stale signals decay; 45 days ≈ the STOCK Act disclosure lag. |
| Net flow | 12% | Buy-vs-sell dollar pressure across all disclosed trades in the name. | Net selling should drag a name down, not just fail to help it. |
The Copy ✔ gates
A high score alone is not enough. A name is flagged Copy ✔ only if all of these hold:
- It resolves to a real, live price (filters junk tickers from annuity/“other” filings).
- Net-positive dollar flow and more buys than sells.
- Score ≥ 55.
- Return since disclosure is KNOWN and ≥ 0 — the copy thesis has actually worked so far.
- Price is KNOWN to be above its 50-day moving average — the move is still alive.
“Known” is deliberate: when a return or the moving average can't be computed, the gate fails — unknown is never treated as good. Real example: NVDA once showed +15% since disclosure but sat below its 50-day average, so it was excluded — the move had already reversed, and copying it would have been chasing a dead trade.
Why entry = disclosure date
Politicians report trades up to ~45 days after making them (STOCK Act). You can only copy a trade once it becomes public, so every return here is measured from the disclosuredate — the price a copier could actually have gotten. Measuring from the politician's own transaction date would flatter the strategy with gains nobody copying it could capture.
Limitations — read these
- The weights are reasoned, not backtested. “It went up since disclosure” is partly hindsight — a name already +20% may be a late entry. The momentum gate mitigates this; it doesn't eliminate it.
- Disclosed amounts are broad ranges (e.g. $1,001–$15,000); we use midpoints, so dollar-flow figures are estimates.
- Every politician is weighted equally for now, even though some historically outperform others — per-politician track records are shown on their profiles but not yet folded into the score.
- Congress discloses up to ~45 days after trading. Everything here is delayed by design; there is no inside track.
Nothing on this site is financial advice. Data is delayed by law and by processing. If you pair a desktop agent, every order it places runs against your own brokerage account, under your sole responsibility.