This seems like a pretty easy question to answer. Just start with Alabama and go from there, right? I thought I would be just a little more scientific, though, and compile actual data By my count, players from over 200 different colleges started NFL games during the 2013 season. Of that group, 37 colleges each accounted for 100 starts or more and, as a group, 57% of total starts.
The following table lists the 37 teams in order of the number of 2013 NFL games started. Those starts are divided into (1) those by players who entered the NFL before 2009 and (2) those who entered the league in 2009 and after. Finally, the table also includes each college’s won/loss record from 2009 through 2013.
While Alabama does rank sixth in number of games started by its former players, the table shows that the answer to the “what colleges” question is not as cut and dry and some may think. The table includes a few surprises both in the omission of some teams I expected to see (e.g., Clemson and Missouri) and the inclusion of some teams I would not have expected to see (e.g., Illinois and Virginia). In fact, about 25% of the teams listed had won/loss records below .500 for the years 2009-2013. California is the poster child for these teams. The college is ranked in a tie for fourth most 2013 games started (and ahead of Alabama and Ohio State) while posting a 24-38 record.
The analysis was then extended to determine whether the 37 colleges were a more efficient source of talent than rest of the college football universe. The term “efficiency” is used in a broad sense to indicate whether the yield (measured by the number of 2013 starts in this case) exceeds what might have been expected given the number of draft selections.
In this article, efficiency is reviewed for each round and the measure used is the number of games started per draftee (“GSD”). The GSD calculation is limited to players drafted between 2009 and 2013 and the games those players started during the 2013 season. GSD cannot exceed 16, the number of regular season games. As an example of the calculation of GSD, the 37 colleges accounted for 117 first round draft selections that started 1244 NFL games in 2013. The GSD would be 10.63, derived by dividing 1244 by 117.
The next table shows a comparison of the 37 colleges to the remaining colleges and then all colleges.
The respective GSDs indicate that, with the exception of sixth round draftees, players from the Top 37 tend to be more successful than draftees from other colleges, meaning those draft choices were more efficiently used.
The analysis was then further extended to review individual colleges within the Top 37. The principal issue in doing so is the limited number of data points for each college. No school had more than 34 draft selections (stretched across seven rounds) and Boston College had only five draftees in the 2009 through 2013 time period.
The next table shows the GSD by round for each the 37 colleges that had at least 20 players drafted from 2009 through 2013.
Where “None” is used, it means a team had no draft selections in those rounds. No attempt is made to present an overall GSD due to the impact that location of draft choices would have on the calculation. Miami, for example, had no first round draft choices from 2009 through 2013, likely resulting in a lower overall GSD than a team like Alabama, which had 14 first round selections in the same period. That lower GSD would be driven largely by the quantity of draft choices and would not be representative of the efficiency of the draft choices that a team did have.
The data presented in the above table can be summarized into which teams were more (and less) efficient by rounds. Generally, three teams are listed by round and category except where there are ties.
It is interesting that Alabama does not appear on the “most efficient” list for any round. Georgia and Florida are the only colleges to appear three times in the most efficient column. Are Alabama players “overdrafted”? This at least raises the issue.
On the least efficient side, Ohio State appears on that list for five of the seven rounds. Another potential case of “overdrafting”? Again, no definitive answer here but the issue is raised.
Follow Tony on Twitter @draftmetrics