A Must Read Statistical Analysis: Do Recruting Rankings Mean Anything?

 Send This Story To A Friend

Everyone is talking about recruiting this week. 

“Is this guy committed?”  “Has this guy wavered?”  “Can we steal this guy?”

And of course, “Where does our class rank?”

For years, I’ve thought that basketball recruiting rankings were a much better judge of future real world success than football recruiting rankings.

For one thing, basketball players are constantly going against each other in summer camps and summer leagues.  In football, it’s more of a guess to compare a kid from North Dakota to another from Florida.

Another reason I think basketball rankings mean more is the fact that an entire team can be greatly impacted by one terrific player.  Carmelo Anthony, for example.  In football, it’s hard to tell how many out of 25 signees will ever make an impact at a school.

Finally, and this is the case with both football and basketball rankings, if you pay close attention you’ll notice that some players go from 3-stars to 4-stars when they sign with a marquee program.  Call it “the Notre Dame factor.”

“Well, if Notre Dame is after him, he must be better than we thought.”

The SEC is one league that benefits greatly from this phenomenon.

Knowing all of that, I still wanted to do a statistical breakdown on recent recruiting classes to see just how accurate they are when it comes to predicting future on-field success.

To do this, I used the rankings from Rivals.com dating back to 2002.  In terms of on-field results, I looked only at the SEC records of the teams in the conference.

Any SEC team should be able to whip a Western Carolina, so I cut out all non-conference games from the study.  That leaves 49 games each year (48 regular season games plus the SEC Championship Game).

By nature, this examination also includes such factors as injuries, player development, transfers, coaching ability, etc.

So what did I find?  Well, there is some correlation between highly ranked recruiting classes and winning… but only in general terms, not in specifics.


First, let’s look at each SEC team’s recruiting ranking from 2002 through 2008, according to Rivals.com.

Let’s now look at those numbers in groups of five.  For example, to judge success on-field, you need to look at all the recruiting classes, not just one or two.

Let’s start with the years 2002 through 2006.  For the 2006 season, this would include everyone from true freshmen to fifth-year seniors.

When you break it down that way, we can take the average five-class ranking from Rivals for each SEC school and compare it to their actual in-conference won-loss record and their recruiting ranking inside the conference:

So what does this show us?  No clear-cut correlation between winning and good recruting rankings. 

Florida had the second best recruiting classes in the SEC (averaging 9th best in the nation) and they had the league’s best record, 8-1 in the SEC (including the SEC title game).

But Georgia had the best five-year recruiting haul (average class was about 6th in the nation for that span) and they finished just 4-4 in the conference… the same record as Kentucky who was 11th in the SEC in recruiting over the same span.

So it’s all meaningless?  Not totally.  Bear with me.



Here are the numbers from 2007, again looking at each school’s average Rivals recruiting ranking (2003, 2004, 2005, 2006 and 2007), their conference record and their recruiting ranking inside the SEC for that five-year span:

So Florida had the best five-year haul and managed only a 5-3 conference record with those recruits.  Again proving that a star-ranking system can’t grade intangibles like toughness and focus.

Mississippi State ranked just 10th in the SEC for the same period and finished just one game behind the Gators in terms of conference record.



But I’m not done yet.  Let’s look at this past season using the same criteria:

Florida had the best classes in the SEC from 2004 to 2008.  They also had the best record in the league and won the SEC title game.

But LSU had the third best talent over that period and they finished just 3-5 in SEC play.



Alright, so why did I say there is some general correlation between these rankings and on-field success… when we can clearly see that there are no specific guarantees?

Look at the teams in terms of groupings:

In 2006, those teams that had the top four SEC recruiting classes from 2002-2006 (Georgia, Florida, LSU and Tennessee) finished with a combined conference record of 23-10… well above .500.

Those teams that finished five through eight in SEC recruiting classes for the five previous signing classes (Auburn, South Carolina, Alabama, and Arkansas) had a combined record of 18-15… about .500.

Those teams that finished nine through twelve in the league in those five recruiting classes (Ole Miss, Miss. State, Kentucky and Vandy) wrapped up the 2006 with a combined record of just 8-24… well below .500.



See a pattern?  It continues:

The top four recruiting schools from 2003-2007 (Florida, Georgia, LSU and Tennessee) were a combined 24-10 in 2007.

Schools five through eight (Auburn, South Carolina, Alabama, and Arkansas) were a combined 16-16.

The bottom four schools in recruiting for that span (Ole Miss, Miss. State, Kentucky and Vandy) finished just 9-23 in league play.



The pattern still holds for the past season, too:

The top four recruiting classes from 2004 through 2008 (Florida, Georgia, LSU and Alabama) resulted in a 25-9 overall record.

The middle of the pack schools (Auburn, Tennessee, South Carolina and Ole Miss) finished 14-18, below .500.

The dregs of the conference (Arkansas, Miss. State, Kentucky and Vandy) finished a combined 10-22.


Conclusions:

1)  You can’t look at a single recruiting class and expect immediate success (or failure) for your favorite team.

2)  You can’t even look at a five-year period (from true freshman through fifth-year seniors) and get a guaranteed indicator of success.

3)  However, if you’re favorite team ranks highly in the SEC over a five-year recruiting span, you can GENERALLY assume that your team will have an above-average season.  The worse you go in overall rank in the conference, the worse you can EXPECT your team’s record to be.  But we’re talking ranges here, you can’t predict specific records with recruiting rankings.



Recruiting rankings are far from definitive.  To be honest, I still believe fans spend too much time studying them. 

But while it’s not a science and it’s certainly not an accurate predictor, it does stand to reason that five good recruiting classes in a row should result in solid on-field results.  Just don’t think that a #1 recruiting class is going to equal a #1 ranking.

In fact, let’s combine all of our data and see how each team has done over the past few years.

Here are the average recruiting rankings from Rivals.com for each team in the SEC from 2002 through 2008 (seven classes) as well as the overall records for those schools from 2006 through 2008 (three seasons, encompassing those seven classes):

Again, it is more telling feature to look at groupings of teams than it is to look at specific schools.

The four schools that had the seven best recruiting classes in the SEC from 2002 through 2008 (Georgia, Florida, LSU and Auburn) finished a combined 66-33 on the field in the last three seasons.  They won about two out of every three games.

The four schools that had classes five through eight (Tennessee, South Carolina, Alabama and Arkansas) combined for a 51-48 record… right about .500.

And the four schools with the worst recruiting classes over that seven-year span (Ole Miss, Mississippi State, Kentucky and Vandy) finished at 30-67.  They lost about two out of every three games (actually a little less).

Even in the big picture, this general look at recruiting seems to back-up our earlier findings.



So come Wednesday, when you’re fretting over your team’s haul, better look at the four previous signing classes, too.  Then compare all those numbers to the other schools in the conference.

That should give you SOME idea as to what you can expect on the field this fall.  Just don’t write it in stone.



UPDATE — Jon Solomon of The Birmingham News (who I like because he’s as big a numbers geek as I am) has now done just about the same study I did.  He’s looked at the national rankings (not just the SEC) and he’s gone back to 2004, rather than to 2002.

His findings are similar to mine.  You can’t draw a guarantee from a recruiting ranking… but if your team does really well for four or five straight classes, you can expect to be pretty good.

Again, keep things general.

As for individual player grades, well, those aren’t as accurate as most would think.

There are just too many variables involved: physical development, coaching, mental development, attitude, behavior, injuries, etc.

Signing a really strong, highly-ranked class just helps your odds of finding some jewels in the haul.  It doesn’t mean that every four- or five-star player will be “a stud” as so many messageboard posters tend to believe.