But it is my goal here to introduce as much order as possible into the chaos that is sports and college football in particular. The challenge I aim to tackle here is can we predict how well a running back will do in an upcoming season using statistical results from past seasons.
My model has five primary variables:
1) Rushing yards
2) Percent of team rushing yards
3) Percent of returning team rushing yards
4) "Depth chart" position (based on yards gained)
The model does not include the potential impact of incoming freshman, returning players at other positions (e.g., OL), or talent evaluations/off-season development, and so is limited compared to a human evaluation. But, like any good statistical model, it has the unique advantage that it can consider historical tendencies across thousands of college football players and dozens of variables simultaneously, something a human evaluation cannot match.
So, drawing on those results, I present the projected top 25 running backs of 2012 by rushing yards. I start today with 25-21 and will continue to present the rest of the top 20 over the next couple of weeks.
. . . . .
25) Zac Stacy - Vanderbilt
Stats: 1193 (2011); 1208 (projected 2012)
The model predicts a slightly more effective Vanderbilt running game, but with the majority of the additional yards going to true sophomore Jerron Seymour.
24) Hunter Lee - Louisiana Tech
Stats: 650 (2011); 1249 (projected 2012)
As a walk-on freshman, it took Lee a while to see much of the field in 2011, but he was a central part of the offense by the end of the season. With Creer moving on to the League, the model sees Lee in a position to have a dark horse, break out year: he is one of the few names on this list that is not on the Maxwell Award watch list.
23) Kenjon Barner - Oregon
Stats: 939 (2011); 1264 (projected 2012)
The Ducks have options in the backfield to pick up the yards vacated by James, but the model sees Barner as the primary beneficiary. Combined, Barner and Thomas should be closer to 2,500 than 2,000 yards.
22) Eddie Lacy - Alabama
Stats: 674 (2011); 1266 (projected 2012)
The model hesitates before projecting double the production from a season ago, but in this case the model is undoubtedly being too conservative. I assume Lacy will be closer to Richardson's 1,679 from a season ago, but the model sees a big drop (about 600 yards) in total production for the Tide from last season.
21) Zach Line - SMU
Stats: 1224 (2011); 1270 (projected 2012)
You know what you are getting in Line and, assuming he can avoid injury and the Gilbert Effect is not overwhelming, Line should be closer to his 2010 mark of 1,494 yards.