Resolved / on record
0/17
No score yet.
17 forecasts on record, none resolved. First result: October 26.
Approved0
Not approved0
Void0
Pending17
Error
- buchiang —no score yet
- Base rate only —no score yet
Mean squared error (Brier score); lower is better. The scale ends at .25, the error of always saying 50%.
Head to head 0 W · 0 L · 0 T
W = the forecast was closer to the outcome than the base rate
Sample0 / 40
At 40 results we call it: if the forecasts cannot beat the base rate, we stop.
Pending · by date
On current forecasts about 12.4 of these 17 get approved; the base rate alone says 13.9.
| Decision | Drug | Forecast | Base | Error if approvedForecastBase | Error if not approvedForecastBase | Left |
|---|---|---|---|---|---|---|
| 10/26 | GSKbepirovirsen | 94% | 95% | forecast .004/base rate .003 (lower) | forecast .884 (lower)/base rate .902 | 21d |
| 10/30 | INOINO-3107 | 40% | 76% | forecast .360/base rate .058 (lower) | forecast .160 (lower)/base rate .578 | 25d |
| 11/01 | AGIOmitapivat | 87% | 95% | forecast .017/base rate .003 (lower) | forecast .757 (lower)/base rate .902 | 27d |
| 11/14 | BTAIQIGALMI | 70% | 95% | forecast .090/base rate .003 (lower) | forecast .490 (lower)/base rate .902 | 40d |
| 11/14 | CYTKaficamten | 96% | 95% | forecast .002 (lower)/base rate .003 | forecast .922/base rate .902 (lower) | 40d |
| 11/14 | SMMTivonescimab | 45% | 76% | forecast .303/base rate .058 (lower) | forecast .203 (lower)/base rate .578 | 40d |
| 11/22 | CAPRderamiocel | 20% | 76% | forecast .640/base rate .058 (lower) | forecast .040 (lower)/base rate .578 | 48d |
| 11/22 | SVRAmolgramostim | 79% | 76% | forecast .044 (lower)/base rate .058 | forecast .624/base rate .578 (lower) | 48d |
| 11/27 | BBIOBBP-418 | 93% | 76% | forecast .005 (lower)/base rate .058 | forecast .865/base rate .578 (lower) | 53d |
| 11/27 | GSKneladalkib | 92% | 76% | forecast .006 (lower)/base rate .058 | forecast .846/base rate .578 (lower) | 53d |
| 11/28 | PYPDD-PLEX100 | 70% | 76% | forecast .090/base rate .058 (lower) | forecast .490 (lower)/base rate .578 | 54d |
| 11/30 | COGTbezuclastinib | 88% | 76% | forecast .014 (lower)/base rate .058 | forecast .774/base rate .578 (lower) | 56d |
| 11/30 | VRTXpovetacicept | 93% | 95% | forecast .005/base rate .003 (lower) | forecast .865 (lower)/base rate .902 | 56d |
| 12/12 | VNDAimsidolimab | 70% | 76% | forecast .090/base rate .058 (lower) | forecast .490 (lower)/base rate .578 | 68d |
| 12/17 | CORTrelacorilant | 40% | 76% | forecast .360/base rate .058 (lower) | forecast .160 (lower)/base rate .578 | 73d |
| 12/22 | MLYSlorundrostat | 88% | 76% | forecast .014 (lower)/base rate .058 | forecast .774/base rate .578 (lower) | 78d |
| 12/27 | PRAXrelutrigine | 80% | 76% | forecast .040 (lower)/base rate .058 | forecast .640/base rate .578 (lower) | 83d |
Under each outcome, two errors: the forecast’s, then the base rate’s. The dot marks the lower one: who would be closer.
All 17 · by date
Seventeen, none developed yet.
Scoring
How it is scored
- Error (Brier score)The squared gap between the forecast and the outcome: approved counts as 1, not approved as 0. Lower is better.
- Against the base rateThe same decisions, scored again as if only the historical base rate had been quoted.
- The two-day ruleOnly the last version recorded at least two days before the PDUFA target date counts; if the FDA decides early, only versions from before the day of its decision. Later changes are not scored.