Navigation: Jump to content areas:


Pro Quality. Fan Perspective.
Login-facebook
Around SBN: 2011 In Extreme Home Runs

Do Recruiting Class Ratings Matter?

Die-hard fans of college football have often heard opposing opinions regarding the ranking of college football recruiting classes: "I know of a player who was a two star recruit, and he's awesome so rankings don't matter!," "Penn State and Ohio State get all the blue chip recruits, that's why they win," "Boise State didn't have any five stars on the team when they beat Oklahoma, so only the coaching matters!," and "USC, Florida, and Alabama will always be in the hunt because they always have highly rated recruiting classes."

Well, what is the actual correlation between recruiting class rankings and results?  For the purpose of exploring this question, let us take a look at the Big Ten in 2008 and 2009.  Of the available recruiting class ranking services, I've chosen to use Rivals.com for the sake of convenience and simplicity.  I used the total class score for the analysis, since it takes into account both the class quality and the class size.  I did not account for recruits who were transfers and non-qualifiers, since I have a full time job and don't have that kind of time on my hands.  While not exactly precise, every team loses players for a variety of reasons. 

First, let's look at the 2008 season since there is a full twelve games to analyze.  The recruiting class scores were statistically weighted for years in the program (seniors have a greater impact than freshmen), and averaged out accordingly for the current Big Ten teams.  Here's what the graph looks like regarding what teams should have the best talent on the field:

 2008bargraph_medium

 As you can see, there appears to have been three distinct tiers of talent in the Big Ten in 2008.  Tier 1: Michigan, Ohio State, Penn State; Tier 2: Michigan State, Iowa, Illinois, Wisconsin, and Purdue; Tier 3: Minnesota, Northwestern, and Indiana.  So, how did these teams perform in relation to their talent levels?:

 2008predicted_medium

With the exceptions of Northwestern and Michigan, every team finished the season within three positions of what their talent levels predicted.  Northwestern and Minnesota were the over-achievers, while Michigan was a catachlismic underachiever.  Granted Michigan was in a transition year, but they only managed three wins despite having some of the most highly rated players in the country.

Let's look at the 2009 predictions:

  2009bargraph_medium

 For 2009 there are still three tiers of talent, with Purdue dropping down to the third tier being the only major shift.  Yes, the 2009 conference schedule is barely under way.  Let's still look at how the teams are stacking up versus this prediction.  I just list the team records, since after two weeks of Big Ten play there isn't much accuracy in the rankings: 

2009predicted_medium

 Taking account that Michigan is still in a "transition" to the spread offense, it is no surprise that Ohio State, Penn State, and Iowa are in the Big Ten title hunt.  Wisconsin, at 5-0 overall, is outperforming expectations so far.  However, they have a tough run of games coming up starting with Ohio State.

The bottom line is that recruiting class rankings have a significant influence upon where a team finishes in the final standings of the conference.  Coaching can on occasion make a big difference in a team's final standings, such as with the 2008 Northwestern and Michigan teams.  Those scenarios appear to be the exception to the rule.  Out-coaching a team that has faster, stronger, and bigger players is a tough mountain to climb.  Based upon this analysis, recruiting class rankings seem to be a very accurate way to predict a team's success.

For those looking to 2010, here's a chart of the recruiting class average star ratings and recruiting class scoring for 2005-2009.  Minnesota fans, take notice of the change in recruiting class scores starting in 2008.  Click on the picture for a larger, more legible view.  With another strong recruiting class in 2010, expect the Gophers to jump up into the second tier of Big Ten talent:

 2009recruitingscores_medium

Comment 22 comments  |  1 recs  | 

Do you like this story?

Comments

Display:

2008

shouldn’t MIST be +1?

"they're calling insane hogs???"

by CrowTrobot on Oct 6, 2009 11:36 AM CDT reply actions  

otherwise

I liked the number crunching, good stats to chew on.

"they're calling insane hogs???"

by CrowTrobot on Oct 6, 2009 11:40 AM CDT up reply actions  

good catch

I’ll correct that when I get a chance

Buck Bravo

by Buck Bravo on Oct 6, 2009 11:48 AM CDT up reply actions  

Great Analysis

Very interesting read. I agree that with another solid recruiting class the gophers could easily move into that second tier of the big 10.

by MinnesotaCatFan on Oct 6, 2009 11:41 AM CDT reply actions  

Rankings by recruiting class

Great graphic depiction of the mountain the lower teirs have to overcome.

by bobbyspringfield on Oct 6, 2009 11:54 AM CDT reply actions  

Im confused

You show Minnesota as the 8th team in the Big 10 for 2008 when rivals had them number 17 in the country for recruiting that year and 3rd in the Big 10. Im not following your thread here. In 09 we also had a higher recruiting class then most the Big 10 outside of the usual Michigan OSU, MSU, PSU i think we were 5th. We have been well ahead of Wisconsin and Iowa in recruiting the last 3 years 08-10.

by Jeffdu on Oct 6, 2009 3:23 PM CDT reply actions  

I think

 that it is not a one-year snapshot. BB is taking four years of recruiting rankings leading up to each season. I’m sure he’ll comment when he gets a chance but that was my assumption.

what you say here can, and will, be used against you

by GopherNation on Oct 6, 2009 3:32 PM CDT up reply actions  

GopherNation is correct,

it is a cumulative weighted average for ALL of the players on the team, not just that year’s recruiting class.

Buck Bravo

by Buck Bravo on Oct 6, 2009 3:35 PM CDT up reply actions  

ratings matter to some extent

There’s a lot of subjective judgment in the “stars” – for example, it seems a recruit’s Rivals star rating goes up the minute he commits to Florida State University.

The “system” will miss on individuals, but there’s strength in numbers & that’s where they’re worthwhile.

But they don’t tell you how good a player is, they’re just a prediction of how good a player might become.

by BD57 on Oct 6, 2009 3:31 PM CDT reply actions  

Great stuff man

Might do something like this for the Pac-10.

Email: bearsnecessities@gmail.com

by Avinash Kunnath on Oct 6, 2009 8:16 PM CDT reply actions  

Query

Does this analysis take transfers and other program attrition into account? Michigan, for example, lost 5* Ryan Mallett, highly rated OG Justin Boren and a host of other players. Their “cumulative star rating” would be a LOT lower if based on actual players vs. classes-as-recruited.

by Mick0212 on Oct 6, 2009 9:39 PM CDT reply actions  

It does not

take transfers and attrition into account. I didn’t have the time to run those numbers manually, team by team. I’d love to see the results of such an endeavor. All teams have some attrition, so it may balance out to some extent. Granted Michigan had quite a bit of attrition for 2008, but their two-deep was probably still more talented than some Big Ten starting lineups. Not to say that RichRod won’t turn things around, but I don’t buy that the 2008 team had the talent level of a three win team.

Buck Bravo

by Buck Bravo on Oct 6, 2009 11:33 PM CDT up reply actions  

Anyone who wants to know

why Michigan had such a terrible year last year has the tools on the Internet to find out why. While a number of reasons had to do with replacing the 2008 No. 1 overall draft pick, 4 year starting QB and RBs, Mario Manningham, Adrian Arrington and 4/5 of their offensive line, there were probably 25 other things that went into it.

That includes Lloyd Carr’s disastrous recruiting and player development near the end of his tenure. Yeah the numbers might be pretty and all. But take his 2005 class, ranked #2 nationally. It had 23 commitments. Of those 23, 6 remain as fifth-year seniors (only 3 of whom start), Mario Manningham, Terrance Taylor, Brandon Harrison, and Carlos Butler left after 2008, and the rest (including 4 four star players and one five star player) either flunked out, transferred, were kicked out of school, or suffered career-ending injuries. To only have 10 players in a class make it through 4 years in your program is devastating to depth and experience. It’s a significant reason why 89 out of Michigan’s current 125 man roster are freshmen and sophomores.

As a point of comparison, Minnesota currently has more senior starters (14) than Michigan has seniors, fifth or fourth year, on its roster (11).

by JG2112 on Oct 7, 2009 6:51 AM CDT up reply actions  

Great work BB

Man, this really drives home the point of how hard it is to climb up a tier.I also like that you used a weighted average because good teams will usually have more seniors and juniors on them. We don’t have very many on our team. I would be very optimistic about the Gophers if it wasn’t for the fact that replacing Decker will be impossible.

by Narby on Oct 6, 2009 9:52 PM CDT reply actions  

Fantastic!

Thanks so much for this BB.

Look forward to your projected 2010 based on this when the recruiting classes are final.

by Django Z on Oct 7, 2009 3:10 AM CDT reply actions  

One request

can you make that last chart larger – it doesn’t get bigger when I click on it and I want to share it with a friend.

by Django Z on Oct 7, 2009 3:11 AM CDT up reply actions  

right click on picture, save it to your computer, then open with Microsoft Picture Manager, use the tools to resize it. This will work if BB can’t fix it.

"they're calling insane hogs???"

by CrowTrobot on Oct 7, 2009 7:56 AM CDT up reply actions  

Nice work. How did you do the "statistical weighting" of the classes?

I am assuming that your “weighting for years in program” is what accounts for the lag between recruiting class and the players actually making it to the field. [I believe this lag to be between 2 and 3 years.] How does it account for players good enough to start as Freshmen (Terrelle Pryor or Tate Forcier, for example)?

I am also looking with interest at Michigan State’s one thousand-point jump from the 2008 to 2009 classes. Do you expect that with another, even stronger class on the horizon for 2010, MSU will jump up into the first tier of talent come 2011 or 2012?

by CPT Hoolie on Oct 7, 2009 8:53 AM CDT reply actions  

For example,

the current freshmen may have half the strength of the current juniors in the weighted average. It’s based upon a few trends I analyzed last year. By no means a perfect Science, but fun to look at and with some obvious relationships between class rankings and standings.

Buck Bravo

by Buck Bravo on Oct 7, 2009 3:44 PM CDT up reply actions  

Ugh!

Neophytes in data analysis! Kill me now!

Correlaton doesn’t prove causation is the first issue and a cardinal sin you have committed in you data analysis, it is merely an indicator that something might be going on…

Second, When you try to show a correlation, how about actually using the two things you’re trying to correlate in the same chart! Take the “Recruit Ranking” as you have it defined here and plot it against something that is a measure of the team’s success. (Not win loss records – too biased in terms of strength of schedule)

Try plotting (for more than the Big Ten, “Recruit Ranking” aganst somethng like the team’s Sagarin Rating! Then you might actually make a better argument.

by St8rBoiInMN on Oct 7, 2009 1:54 PM CDT reply actions  

It's not

supposed to be a statistical study. It’s just a view of relationships. Sure, it’s possible that it is all due to random chance. But most people don’t want to read about p values and Student T tests. I can do all of that if you pay me, but I don’t have the time and most of our readers don’t want to read it. Your tone aside, you are technically right.

Buck Bravo

by Buck Bravo on Oct 7, 2009 2:24 PM CDT up reply actions  

Buck

your points are well taken and reinforces what we probably instinctlvely knew. And true, we are not interested in squabbling over whether inferential statistics or mathmatical analysis should be used.

by Texas Gopher on Oct 8, 2009 11:54 AM CDT reply actions  

Comments For This Post Are Closed


User Tools

Minnesota Golden Gophers sports--Usually analytical, sometimes snarky.

FanPosts

Community blog posts and discussion.

Recent FanPosts

Epson003-2_small
Gophers @ Nebraska
47487
New Uniforms!
M_nerf_small
B1G/PAC Matchups
Small
Can Minnesota get to a bowl game in 2012?
Ishot-2_small
A modest proposal to eliminate 15 bowl games and send their mediocre teams home for the holidays
Mn_basketball_small
Big Ten Bloggers Pick'em Standings
Small
Minnesota Gophers Should be Happy to Have Jerry Kill
Small
Eat More (Bucknell) Bison

+ New FanPost All FanPosts >


Managers

Mn_basketball_small GopherNation

Imagescary7xfo_small Jeffrick

Img-20101223-00043_small JDMill

Editors

Imgres_small PJS

Bigbuck_small Buck Bravo

Dsc02076_small JG2112