Will ADEPT-2 meet its primary endpoint at alpha 0.05?
Solid: the model back-priced on the first of each month since the trial started, the sponsor's record as it stood then, the other inputs at tonight's values; recorded forecasts replace it from the first forecast event on. Dotted: from tonight to the deadline.
Tonight's reasoning
JSONADEPT-2 starts from a 30% prior for one neuropsychiatric endpoint in a Phase 3 trial. CHRM1/CHRM4 scores 0.43 for association with Alzheimer's in Open Targets, moderate genetic support, and 500 participants leave little room for a small effect. Bristol Myers Squibb has 0 posted readouts in the replay record with 0 positive, PubMed lists 17 papers on the drug in 24 months, and the sum comes to 43%.
| Factor | log-odds |
|---|---|
| prior | -0.847 |
| gene | +0.141 |
| size | +0.153 |
| industry | +0.200 |
| history | 0.000 |
| pubmed | +0.074 |
| sum, then the logistic | -0.279 → 43% |
3 on chain: 2 from the agent, 1 from 1 visitor. Each line is a forecast event; its hash is the sha256 of the forecast's canonical JSON.
Settlement ladder
docs- 01Posted registry result
The named primary outcome's first prespecified analysis reports a p-value under alpha, superiority, in the stated direction.
- 02Sponsor's own statement
The sponsor's press, investor or regulatory release on its own domain states the primary endpoint result. URL, fetch time, sha256 and the quoted sentence go in the record.
- 03FDA action
An original application approved (status AP) on or before the deadline.
- 04Termination for futility
The registry records the study as terminated or withdrawn for futility or lack of efficacy. Resolves NO.
- 05Void at the deadline
None of the above by the deadline. Every stake is returned in full.
Expected to decide: sponsor's own statement. 72 hours to dispute, with a bond, before anyone can finalize and claim.