> > >
The first cohort scored on the recalibrated engine. Fourteen players who were still in the minors as of June 11 with no MLB time this cycle. Every probability below was locked in public on the run date. The window closes September 9.
| Player | Org | Lvl | Pos | CUS | CUP | PPI | Path | ECD Mid | Flags |
|---|
CUP is a binary 90-day call-up probability. A score of 62 percent means the model gives Thomas White a 62 percent chance of a major league appearance before September 9. It is not a talent grade and it is not a prediction that any single event will happen. The methodology behind the CUP score and the v0.5 mapping change is public.
Most of the preseason Top 100 has already graduated to the majors, so this group skews toward upper-minors arms and the next wave of bats. The three Triple-A bats in the middle, Owen Caissie, Emmanuel Rodriguez, and Kaelen Culpepper, all sit at 33 percent on real organizational runways. The Twins cleared their infield at the deadline, and Miami keeps promoting. Org promotion behavior is doing real work in these numbers.
The speculative group at 22 to 23 percent is mostly Double-A talent, where the level itself caps a 90-day window. Walker Jenkins would have scored higher, but a shoulder injury reported in mid-June applies the injury-severity humility band, the direct lesson from the Aidan Miller miss at the last checkpoint. Two carryovers from the first cohort are scored fresh on v0.5 here: read the original Thomas White and Charlie Condon write-ups for the full context.
The recalibrated CUS-to-CUP table, rebuilt from the 28 resolved Sub-Batch A outcomes:
| CUS band | CUP |
|---|---|
| under 30 | 5% |
| 30 to 39 | 12% |
| 40 to 47 | 22% |
| 48 to 54 | 33% |
| 55 to 61 | 45% |
| 62 to 69 | 62% |
| 70 to 77 | 78% |
| 78 and up | 88% |
On the Batch A cohort, this cut the Brier score from 0.269 to 0.142 against the same predictions. That recalculation is a corrected score on resolved data, not a new prediction. The original 0.269 stays on the calibration log as the honest first result.
One email when the cohort resolves and the Brier score lands.