This is the second diary in a series where I attempt to identify the correlates of recruiting and test the strength of factors we often think to be relevant for recruiting success. The first diary discussing and describing the data can be found here. 

In this installment, I look at two factors that are often brought up in discussions of recruiting: geography and recent NFL draft picks. Do these factors matter in recruiting elite (top-200) prospects and, if so, how much of an impact do they have?

Key takeaways (tl;dr) 

1. Geography matters. Not surprisingly, schools that are in the same region/conference footprint as the recruit have an advantage in gaining a commitment. This finding was pretty consistent across regions and conferences (i.e., SEC schools have an advantage with recruits in the south/SEC footprint, B1G schools have an advantage with recruits in the Midwest/B1G footprint, etc.). But, and this is a BIG but(t), the heavy concentration of top recruits in the south/SEC territory means that schools like Michigan will always have a difficult time consistently pulling in top-200 recruits.
2. Geography matters…unless you are Alabama or Georgia. When Saban or Smart come calling, recruits listen. Even if they reside outside of the South or SEC territory.
3. Draft picks matter, to an extent. As we would expect, recruits are more likely to commit to a school with a greater number of draft picks in the recent past. Also as expected, a greater number of early round picks as well as draft picks at the recruit’s position are even more relevant. But, and again this is a BIG but(t), unless you are putting an obscene number of players in the pros, recent draft success is not a panacea for geographical disadvantages. 
4. Conference realignment and getting a footprint in California (and potentially Florida) may be the best chance that a school like Michigan has to consistently pull in elite recruits.

THE DATA

The base data include the 247 composite list of top-200 players in each recruiting class for the past 5 years (2019-2023).The data are organized by the recruit and the offer school. That is, for each top-200 recruit, there is a separate observation (row) for each P5 school that made a scholarship offer as listed on each recruit’s profile page on 247sports.com. For example:

Organizing the data in this way allows me to test how characteristics of the different schools (geography, recent draft picks, recent on-field performance, etc.) influence a recruit’s decision to commit to one team over another. Since the dependent variable in all models is dichotomous—a recruit either commits to the school (1), or doesn’t (0)—all models are estimated using Probit.

THE RESULTS

Hypothesis 1: Geography matters

How much does a school’s location in the same or different geographic region/conference footprint as the recruit matter for recruiting success? 

To test this hypothesis I coded each recruit and offer school into one of four mutually exclusive geographic regions. Regions are coded according to the US Census Bureau list of regions as follows:

For anyone that is interested, the results were obtained from a probit model interacting the offer team’s region with the recruit’s region: commit(0/1) = β1(offer school’s region) + β2(recruit’s region) + β3(offer school’s region x recruit’s region). For ease of interpretation, I show the graphs of the predicted probabilities (with 95% confidence intervals). I can talk more about the estimation strategies in the comments section if you have any questions.

Unsurprisingly (you’re probably going to hear this a lot), the results show that geography does indeed matter. Recruits are more likely to commit to schools within their home region. What is surprising, however, is that this regional pull appears to be the strongest for midwestern recruits. A team from the midwest has about an 8% chance of receiving a commitment from a midwestern recruit, compared to about a 2% chance for a team from any other region. Southern recruits also have a strong pull to southern teams, though the difference is slightly less pronounced.

The two exceptions are for recruits from the Northeast and West. While northeastern and western schools have an advantage over recruits in their regions, this advantage is much less significant. Notably, midwestern schools have been able to successfully recruit these two regions. In the case of western recruits, this might bode well for schools like Michigan once USC and UCLA join the B1G and become part of the conference footprint. 

In the next models, I break the offer schools and recruits down by their conference footprint. Note, however, than conference footprints, unlike regions, are not mutually exclusive. That is, a recruit from Florida resides within both the SEC and ACC footprints, and a recruit from Texas is in both the SEC and Big 12 footprints. This needs to be kept in mind when looking at the results.

Again, recruits are more likely to commit to schools within their same conference footprint. B1G schools have probably the largest advantage over recruits within the B1G footprint, followed by SEC schools for recruits within the SEC footprint.The strength of the SEC schools for recruits in the ACC and Big 12 footprint is probably a result of overlapping footprints (i.e., in states like Florida and Texas). 

The Verdict: Hypothesis supported

Hypothesis 2: Geography matters, unless you are Georgia or Alabama

Do the same restraints of geography matter for recruiting powerhouses like Georgia and Alabama? 

For this test, I simply removed Georgia and Alabama from the SEC and created a separate variable indicating whether these schools extended an offer to the recruit. The results?

It’s pretty clear that when Saban or Smart extend an offer to a recruit in the SEC or ACC footprint, they are almost always in contention to get a commitment. In both cases, they have an almost 17% chance of signing the recruit, much greater than the rest of the SEC schools (for recruits in SEC territory) or the ACC schools (for recruits in the ACC footprint). They are also very strong contenders for recruits in the B1G, Big 12, and Pac 12 footprints. The results are pretty clear: when Saban or Smart make a call, recruits listen.

The Verdict: Hypothesis supported

Hypothesis 3: Get me to the pros

Are recruits more likely to commit to schools with a greater number of recent NFL draft picks?

For the next set of tests, I created variables indicating the total number of draft picks from each offer school in the previous 2 NFL drafts. It seems reasonable to expect that recruits not only pay attention to overall draft picks, but also the round in which players were drafted as well as the draft picks at their own position. I therefore created four separate variables: 

• # of Draft Picks (Any Round)
• # of Round 1 Draft Picks
• # of Round 1 or 2 Draft Picks
• # of Draft Picks at the recruit's position 

No real surprises here as draft picks matter for recruiting success. It is also not surprising that the strength of the association increases for draft picks in the earlier rounds and draft picks at the recruit’s position (note the steeper slopes for these variables). 

The Verdict: Hypothesis supported

Hypothesis 4: Draft success can overcome geographical barriers to recruiting

We’ve already seen that geography plays a large role in recruiting. But can NFL draft success help schools to overcome geographic disadvantages in recruiting? 

For these tests, I interacted each of the draft pick variables with a dummy variable indicating whether the recruit was in the same (1) or different (0) conference footprint as the offer school: commit(0/1) = β1(# of draft picks) + β2(in/out-of-footprint recruit) + β3(# of draft picks x in/out-of-footprint recruit). 

The number of draft picks certainly matters, but it seems to be more of a factor when deciding between schools within the recruit’s footprint. In general, however, the effect is rather marginal for recruits outside of the offer school’s footprint. 

To see this, note the horizontal dashed line showing the average (unconditional) probability that a recruit will commit to an out-of-footprint school (pr = .026). Therefore, to provide an added advantage, the # of NFL draft picks would have to lead to a probability of commitment higher than this (unconditional) probability of commitment. At 10 draft picks over the previous 2 years, we see an added advantage, but only a marginal one (pr = .028). 

Moreover, to simply match the (average) probability of a commitment to an in-footprint school (pr = .067), a school from outside the recruit’s footprint would need to have at least 18 recruits drafted over the previous 2 years (pr = .068). Needless to say, this is pretty rare (see the underlaid histogram of the statistical distribution of draft picks). 

Does it make a difference if we look at 1st or 2nd round draft picks instead of any draft picks? Not really.

Again, to even match the (average) advantage that in-footprint schools have, an out-of-footprint school would need to have at least 8 1st or 2nd round draft picks over the previous 2 years, something that only occurs in about 4% of the cases in the data.

The Verdict: Hypothesis unsupported, or at least statistically improbable

CONCLUSION

Geography matters in recruiting. And unless a school is putting an absurd number of players into the NFL, it is very difficult to consistently pull top-200 recruits from outside their home region/conference footprint. A simple look at the recruiting map will indicate that Michigan and other teams from the north are at a disadvantage in this respect.

But there may be a glimmer of hope. It seems that The Powers That Be in the B1G may have been onto something with this whole conference expansion thing. Expanding the B1G footprint to California (and, if some rumblings turn out to be true, potentially Florida) would be one way, statistically supported by these data, to get schools like Michigan into some very fertile recruiting territory.

To use a (very stretched) analogy: many of us can remember when Russia/USSR was the predominant recruiting power in Eastern Europe until the end of the Cold War, right? Well, maybe conference realignment will be the first cracks in the foundation of this metaphorical Berlin Wall. And we all know what happened next…

Germany won! (Shit, that can’t be right, can it?) 

In the next installment, the effect of on-field performance (wins, playoff appearances, position-level statistics, etc.) on recruiting success.

Hope you enjoyed. Go Blue!