Source. The explanation below is partly Bill C's, partly my interpretation from years of reading his work and listening to his podcast(s).
For those who aren't college analytics nerds:
SP+ is a tempo- and opponent-adjusted measure of college football efficiency. It is a predictive measure of the most sustainable and predictable aspects of football [meaning it does not consider things like turnovers, penalties, etc]. It is not a résumé ranking, and, along those same lines, these projections aren't intended to be a guess at what the AP Top 25 will look like at the end of the year.
Preseason projections are based on three factors.
- Returning production - how many players are returning, how many plays each one has been a part of, and how successful each play has been (Bill has often said that not all experience is good experience, so it's important to not reward teams/players for participating in unsuccessful plays).
- Recent recruiting - Bill says that this piece makes up about one-fifth of the projections formula.
- Recent history - I think this piece is particularly interesting. Per Bill:
Using a sliver of information from previous seasons (two to four years ago) gives us a good measure of overall program health. It stands to reason that a team that has played well for one year is less likely to duplicate that effort than a team that has been good for years on end (and vice versa), right? This is a minor piece of the puzzle, but the projections are better with it than without.
Anyways, VT at 61....
Pry has his work cut out for him (not that we needed a fancy algorithm to tell us that)