Verdict: Research only. Three seasons of history do not support the idea. We are still tracking it through the 2026 season before we close the book on it.
The question
College football lines open on Sunday afternoon. Most power ratings, ours included, update a day or two later. Sometimes the market has already moved toward where our number ends up. The idea: when the early move and our number agree, the line keeps moving that way, and the bet is better than an ordinary pick.
How we tested it
We pulled the lines for every FBS game from 2023 through 2025 at four times each week: the Sunday opener, Monday at noon, Monday afternoon and Saturday morning. Then we priced every game with a model built only from information available before that week, so it never saw the result, and compared games where the early move agreed with the model to games where it did not.
What we found
- Agreement did not help. Games where the early move agreed with the model covered 50.6%. Games with no early move covered 48.8%, and games where the market moved the other way covered 50.4%. Those are all coin flips.
- The pattern even reversed in 2025. Agreeing games covered 45.3%, disagreeing games 55.9%.
- There is a trap in looking at picks only. Most picks happen when the market moves away from the model, because that is what makes the gap big enough to be a pick. Looking at picks alone makes early moves look more important than they are, so we tested every game instead.
What we are doing with it
Nothing changes on the card. The 2026 season is the real live test, with lines captured every hour, and we will judge it when the season ends. Until then, early movement is something we track, not something we bet.