Track record Does any of this actually work?
Most congressional-trading sites sell a story: follow Congress, beat the market. We test that story against our own data, in public, and publish the result here — updated automatically with every build. Where our signal is just noise, this page says so. Every figure is shown with its sample size (n), because a small sample is how hype hides. Last updated: 2026-11-05 00:32 UTC. New here? Start with can you get rich copying Congress? Scored with Interest Score v1.1 (changelog).
1. Does a high Interest Score predict market-beating returns?
No — and that's the honest answer. A higher Interest Score does not reliably line up with higher excess returns. That's exactly what we'd expect: the score is an attention filter, built to flag which trades deserve scrutiny (size, rarity, clustering, oversight overlap), not to predict which stocks go up. We'd rather tell you that than sell you a signal that isn't there.
| Interest Score band | Avg excess vs S&P | 95% CI | n |
|---|---|---|---|
| 0–20 | -0.9% | -1.7…-0.2 | 1531 |
| 20–40 | +0.0% | -0.8…+0.9 | 1188 |
| 40–60 | +2.0% | +0.8…+3.2 | 755 |
| 60–80 | -0.6% | -1.4…+0.2 | 1501 |
| 80–100 | -0.0% | -0.8…+0.7 | 1545 |
| All buys (baseline) | -0.1% | -0.5…+0.2 | 6520 |
Full method and both perspectives (member vs follower) on the “Does it work?” page.
2. Do members really trade in the sectors they oversee — more than chance?
Here the answer is yes, clearly. We shuffle the member↔trade links 5,000 times to build a “pure chance” baseline. In reality, 8.85% of trades fall inside a sector the trading member oversees; under chance it's 10.71%. The gap is statistically not significant (p = 1.000), across n=19439 informative trades.
This is the accountability pattern the platform exists to surface — and unlike “beat the market,” it holds up. Full permutation test on the “Is the pattern real?” page.
3. When we flag a signal, how often is it right?
Of 662 signals old enough to judge (a 21-day horizon), 50% moved in the called direction versus the market, with an average excess of +1.3% in the direction of the call. Another 478 are still aging and aren't counted yet. A coin flip is 50% — read this number against that, and against the sample size, not against a cherry-picked winner.
Every judged signal, hits and misses, on the signal scorecard.
4. How have high-interest trades done vs the market lately?
High-interest trades are those scoring ≥ 60, entered on their filing date (what a follower could actually do) and measured over 63 days vs the S&P.
n=793
n=1281
Short windows swing on a handful of trades — that's why every bar carries its n. Read the trend, not any single month.
5. Does the abnormal return persist after disclosure? (event study)
A single horizon can flatter or flatter-to-deceive. This event study measures the average market-adjusted abnormal return (vs the S&P) of disclosed buys at several horizons after the filing date — the realistic, look-ahead-free entry — for high-interest trades (score ≥ 60) against all buys. A persistent, significant positive figure would be real evidence of an edge; a flat or inconsistent one is the honest signature of an attention filter.
| Horizon after disclosure | High-interest — avg abnormal return | n | All buys | n |
|---|---|---|---|---|
| 21 days | +0.36% sig (+0.0…+0.7) | 3076 | +0.18% | 6507 |
| 42 days | +0.14% (-0.3…+0.6) | 3008 | -0.08% | 6396 |
| 63 days | -0.32% (-0.9…+0.2) | 2982 | -0.17% | 6311 |
sig = the 95% confidence interval excludes zero at that horizon. Abnormal return = the stock's return minus the S&P 500's over the same window. Read the row-to-row consistency, not any single cell.
Not investment advice. Every figure here is a backtested, hypothetical calculation on public disclosure data — not actual returns, and not a recommendation. Excess return means return above the S&P 500 over a fixed horizon. The point of this page is the opposite of a tip sheet: it's to show, honestly and continuously, what the data does and does not support.