We Tested Whether Congress Trades + a High Score Beat Either Signal Alone. They Don't.
Every number in this piece is from our own backtest archive (data/confluence-backtest.json, 72 real Congress-disclosed trades matched against our own score history), not external reporting. The sample is small — 72 trades, 28 of them landing on a Buy Zone score — and the data file itself says so in its own words: "directional evidence, not statistical proof." We're reporting it exactly that plainly, because the honest answer here undercuts a feature we built.
The idea we tested
A member of Congress discloses a stock purchase. Separately, our own entry score says that same stock is sitting in a clean Buy Zone (score 79+) on the day of the trade. Intuition says combining two independent bullish signals should beat either one alone — a real person with (allegedly) better information, confirmed by a real technical setup. We built a "confluence" check specifically to test whether that intuition holds up against our own numbers.
What the data actually says
It doesn't hold up. Trades where a Congress purchase landed on a Buy Zone score averaged +4.35% since the trade, with a 60.7% win rate (n=28). Trades where Congress bought but the score was below 79 — no technical confluence at all — averaged +8.74%, with a 68.2% win rate (n=44). Both the average return and the win rate point the same direction, and it's the opposite of what the "two signals are better than one" story predicts.
| Confluence (score ≥79) | No confluence (score <79) | |
|---|---|---|
| Trades | 28 | 44 |
| Avg return since trade | +4.35% | +8.74% |
| Avg excess return | +2.77% | +6.09% |
| Win rate | 60.7% | 68.2% |
Why this probably isn't what it looks like
The honest read isn't "Congress trades are a contrarian signal when our score agrees with them." The more likely explanation is survivorship in the setup itself: a stock already sitting at score 79+ has, by definition, already pulled back into its entry zone — there's less room left for the kind of sharp continuation move that drives an outsized return in a short window. A trade landing on a lower score has more room to run if the stock recovers, which inflates the average on the no-confluence side mechanically, independent of whether Congress's information was any good. We can't fully separate that effect from a genuine signal with 72 trades — which is exactly why the underlying file calls this directional evidence, not proof.
What we're actually doing with this
Nothing, for now — and that's the point of publishing it. We built a hypothesis, tested it against our own real data, and the data didn't cooperate. No confluence badge shipped to the Rankings table on the back of this; a 72-trade sample with a plausible confound isn't a foundation to build a UI feature on, regardless of which direction the headline number points. If the sample grows and the pattern either strengthens or reverses, we'll say so — the same way we're saying so now.
Scores in this piece come from the ZonedInOut model. How the score works →
Frequently asked
Does it help when a Congress stock purchase also has a high entry-timing score?
Why would a lower-score Congress trade outperform a high-score one?
Did ZonedInOut add a Congress-confluence badge to the Rankings table?
How many trades were in this backtest?
Get the Sunday Brief
One email a week, before the open: what the board did, which calls worked, which didn't, and what changed. Free. No spam, unsubscribe anytime.