> library("chan.preseason")
> projections %>% arrange(desc(proj_adjem))
Known sources of inconsistency
Non-D1 arrivals still being added. D2, JUCO, and international prospects arrive without matching records in the Bayesian projection pipeline (which is trained on D1-only season stats). These players aren't reflected in the projected roster, so teams relying on them will appear thinner than they'll actually be on opening night. Incomplete rosters distort minutes allocation. The 200-minute allocation is distributed only across known players. A team still finalizing its portal class or freshman signings will have those 200 minutes spread over fewer names, which inflates the allocated MPG for each included player and makes the weighted-average BPM unrepresentative. Projections for teams mid-cycle on roster construction should be treated as provisional.
## Loading...
Rk Team Conf Proj AdjEM Cur AdjEM Cur Rk Rk Chg EM Chg Ret Cont% Key Loss Key Addition