College Football Recruiting Prediction Model

The Stetson School of Business and Economics at Mercer University provides on its website a model which it claims was developed in order "to capture, to the best extent possible, a statistical equation to capture the decision making process" of the nation's top-rated student-athletes coming out of high school with the intention of playing football at the collegiate level.

Noting the significance of acquiring a competitive recruiting class in terms of revenue which could be generated in the future, researchers involved with the creation of the College Football Recruiting Prediction Model -- or CFRPM, for short -- built a database whose aim was evaluating the characteristics and decisions of nearly 3400 high school football recruits comprising the recruiting classes of 2002, 2003, and 2004. Using a special form of what the Mercer website refers to as a probit -- or generalized linear -- model, and recognizing that most recruits generally choose from a grouping of four schools in making their final commitment to a particular program, researchers aimed to create an econometric system -- mathematical and statistical techniques used to predict solutions to specific economic problems -- that could help athletic departments and their corresponding football programs at the NCAA level more accurately predict what a particular recruit might -- or, in fact, might not -- do in terms of making a formal decision to play at one school versus another.

The results which the CFRPM study yielded, says the article, were surprising. Data suggested that factors such as an university's graduation rate, the number of Bowl Championship Series bowl appearance, the current depth at the chosen position for the respective recruit, the number of players drafted by the NFL from a particular program, and even the amount of national championship titles a program achieved did not "systematically influence" the decisions made by the majority of high school student-athletes playing football. Instead, the dynamics most readily identified as having an impact on the choice of one program over another included whether or not the athlete made an official visit to the college selected, whether the school was a member of a BCS conference, the distance from the high school athlete's hometown to the university chosen, whether the recruit came from the same state as the program who recruited him, the final AP ranking achieved in the previous year of competition, whether or not a program was under investigation or penalty as a result of violating NCAA rules, the size of a particular team's stadium, and the overall age and location (on-campus / off-campus) of the program's venue.

The study concludes that "in a nutshell, high school athletes prefer winning programs that are close to home, are in possession of good physical facilities, and are in good graces with the NCAA." Ironically enough, however, reduced scholarships too, notes the study's findings, actually increased the likelihood of a program being selected due to the fact that such a development implied "reduced competition for exposure and playing time in the future."

It is interesting to note that when applied to the top 250 recruits of 2009 (as determined, the article states, by Rivals.com), the CFRPM and associated findings produced a 71% accuracy rate in terms of identifying which schools prospective recruits ended up; out of 245 predictions made, the model ended up correctly identifying 174 of them.

When applied to previous years as well, success rates of between 68% and 73% of predictions proved accurate, with the latter rating coming from the use of the CFRPM on the recruiting class for 2008.