Scuzz Model: Recruiting Rankings (Take 1)
After a healthy break to recooperate from bowl season, watch some Rugby, and deal with winter by gorging during the Super Bowl, the Scuzz Model is back with some analysis on recruiting ratings.
I’ve alluded to this on the podcast a couple times in the last month — essentially I pulled down 10 years of recruiting ratings from both Scout and Rivals, both to assess how recruiting changes over time at a school and to compare the ratings against performance results (both from the Scuzz Model and elsewhere).
Note, all the stats below are based on the average star rating of each recruit — not on the class ranking that Scout or Rivals assigns to each school.
Northwestern vs. the Big Ten
The charts below show how NU’s average recruiting rating over time compares to the average for the other 11 Big Ten schools (Nebraska is included throughout as a Big Ten school – not just in 2011). The shorter lines show a 5-year rolling average of NU’s recruits, or Rolling Recruiting Rating (RRR). This both smooths the raw annual ratings and illustrates the talant amassed in any particular season.
Immediately, you can see NU’s average star rating has risen over time. You can also see that the RRR is increasing faster than the Big Ten average. We have all taken note of the improved talent Fitz is bringing to NU, but to see the gap versus the rest of the Big Ten closing is a really great validation of this story.
What’s also clear (tho moreso for Scout than Rivals) is that the last three years (’10, ’11, ’12) show a real jump in the annual rating, which is of course driving the upward trend of the RRR. This would seem to solve the chicken-or-egg conundrum of college recruiting: clearly these ratings went up because of the Cats’ success in ’08 and ’09. However, consider some of the other Big Ten teams over the same period (Rivals results only):
Some of these results are similar to NU — the RRR for Michigan and Penn State have both tailed off in recent years, while Michigan State has shot upwards. Iowa seems to be languishing, despite having a slightly better record than NU since 2007. And most surprising are the increases at Minnesota and Indiana! We all know there are no wins to back up or explain either of those increases.
To this point, I think there’s something to Fitz’s dismissal of recruiting rankings. There are so many other factors at play… for example, Minnesota hired Tim Brewster who was a dynamite recruiter, but a terrible coach; Brady Hoke inherited diminishing recruiting returns at Michigan and went to a BCS bowl in year one.
Production vs Recruit Rating:
There are a million other analyses out there linking recruiting ratings to future performance — no matter how one looks at it, there are always exceptions – big time players that nobody knew about in HS, big time recruits who never saw the field, teams with great recruiting records who bomb, and teams who greatly exceed their talent rating. My first take on this comparison shows a stronger relationship between performance and subsequent recruiting than recruiting and future performance.
The charts with red data-points show how the Scuzz model’s expected winning percentage during a season correlate to the raw recruiting ratings on the following signing day (i.e. 2011/12 season –> 2012 signing day). The blue charts show how the RRR correlates to performance in the 5th year of that average measure. The red-data shows a much stronger correlation, represented by the r-squared value on each chart. At this point, these data are for the Big Ten only, but I will be expanding this analysis to the BCS in future posts,. Based on this first cut, it seems that performance drives recruiting, more than the other way around.
5 Year Big Ten Breakdown
The last exhibit I’ll share shows how each individual team’s RRR compares to their 5 year on-field performance. I can’t easily combine the Scuzz model’s results across years, so I’ve used the F+ ratings recently released by Football Outsiders. What’s interesting here is to examine the teams that deviate most from the expected relationship between recruiting and performance:
Essentially, the teams above / to the left of the trend-line have underperformed based on the recruiting ratings, while the teams below and to the right have over performed. We can easily point to some examples — Michigan stands out as having the 2nd highest RRR, but falls in the middle of the pack for performance. Proof of both the amazing advantage some schools have in recruiting by way of who they are, and that one terrible coach and a major change in scheme can totally gutter a powerhouse program.
Not surprisingly we find NU in the “overacheiver” bucket, but I am shocked to see the Illini there as well given what we knew about Zook as a recruiter (good) and coach (awful). Iowa & Wisconsin are both teams that recruit more to their scheme/style than most, and have outperformed their perceived talent level the most.
Most interesting to me is Ohio State, who have by all measures dominated the Big Ten in the 5 years represented by these data, but do not live up to the rating of their recruiting talent. The Buckeyes are perhaps hampered more than other Big Ten schools by early departures to the NFL and (ahem) otherwise.