Statistical Models Preview Alabama vs. Texas A&M football
Texas A&M football vs. Alabama is one of the biggest games on TV this weekend, so naturally, many predictions have come in about who will win, who will cover, and everything else having to do with the game. What might the different advanced statistical models make of this matchup? It is, of course, exceedingly rare for Alabama to be predicted to lose based solely on the advanced metrics. So how are the Aggies projected to fare against the Tide? Let’s dive in.
SP+: Texas A&M football vs. Alabama almost too close to call
Bill Connelly of ESPN’s SP+ model is one that I consider extremely methodologically sound. Especially nearer to the beginning of the season when compared to other models, Connelly’s system excels. His is based largely on success rate and a little less so on explosiveness—you can read more about it here. His model sees this game as essentially too close to call, though it has Alabama scoring about 0.5 points more on average.
This essentially has the game as a complete coin flip. Considering preseason priors are still calculated in this formula, and given where each team was in those preseason rankings, I feel good about the Aggies being able to outperform this prediction.
EPA per CFB-Graphs: Texas A&M Football Wins Straight Up
Parker Fleming of CFB Graphs has a rosier outlook for the Aggies per his Expected Points Added model. EPA is a more one-size-fits-all statistic that converts yards gained or lost, contextualizing with field position as well as down and distance, into points. Here’s what Fleming’s model says:
There’s a lot here, but, as he said on his page, this could be the first time the numbers have had Alabama losing straight up. That’s pretty notable. One of the biggest edges the Aggies seem to have here are the ability to prevent Alabama from creating quality possessions (Eckel Rate and Pts/Eckel here), so what happens on the quality possessions that Bama does get will be pretty key.
Points Per Drive Model has Alabama Beating Texas A&M Football
Brian Fremeau of bcftoys.com runs a model based on net points per drive adjusted for opponent quality. His model sees this game as a bit less competitive, projecting Alabama to win by a mark of roughly 24-18. I can’t help but feel as though the turnovers in the last game hurt the Aggies when it comes to their offensive points per drive rating, but in any case, Fremeau’s model sees the Aggies losing this weekend.
All in all, this is a bit down the middle—something that can’t be said for the advanced projections in any other matchup the Aggies have had against the Tide since joining the conference. One model has it as a virtual tie, one favors the Aggies by about 5, and on favors the Tide by about 6. To me, that says we have an exciting game coming.