Event-study prior for 'congress_disclosure' claims points bearish (hit rate 0.62). Congress: 2 buys / 3 sells → bearish. Technicals (yahoo-finance): last 147.64, RSI 56.5286, neutral → bullish. Fundamentals (SEC FY2026): revenue growth 0.0326, margin 0.1844 → neutral. Blended: bearish (confidence 0.57).
Weight answers "how much should this filing matter?" Direction confidence answers "which way, and how sure?" Averaging them would hide exactly the disagreement you need to see.
| Input | Provenance | Pull on signal |
|---|---|---|
| Congress flow: 2 buys / 3 sells (5 recent trades in PG) STOCK Act filings, official clerk records · 30% of blend | REAL | + direction |
| Technicals: last close 147.64, RSI 57, trend neutral yahoo-finance price data · 10% of blend | REAL | against direction |
| Event-study prior for "congress_disclosure": hit rate 0.62, moderate magnitude Hand-seeded prior, not yet measured · 55% of blend | ESTIMATE | + direction (discounted) |
Bracket sums are shown as filed ranges, never point estimates. Paper-filed PTRs that can't be parsed are skipped, not guessed. · Read the official source document →
The adversarial desk review arrives in Milestone 3. Future dockets will publish what was challenged before publication, what the desk conceded, and what it defended. Until then this section stays honestly empty rather than invented.
An opposite filing by the same member in this name, an amendment from the clerk, or the measured event study coming back flat for this filing profile. If any of those land, this docket gets amended in place · with the change logged, not overwritten.
This is a probabilistic read on a public filing, not advice. The 57% means we expect the direction call to be wrong about 43% of the time.