Or: Michigan State Recruiting is Still More-er Than Fine-er: A Second Look at Player Development
Author's note: I owe a giant thank you to Hey Joe! poster '89 Chemistry, who helped out with a large chunk of the player data for a handful of these teams. It was a big help.
Also, Longpost is Long. Like, so long.
When I wrote this piece last year, many Spartan fans seemed to be angry/depressed that recruiting was going so slowly compared to some of our rivals. 11 wins, a good 2012 class, and a fairly fast and strong start to the 2013 class later, and many of those same fans have found a way to take the ups and downs of recruiting in stride. Rather than fretting about how we aren't going to be able to compete unless we get a bunch of 4 and 5 star talent, the general consensus of fan base has basically shifted to "Dantonio gonna Dantonio, and that's probably going to work out pretty well". So, excellent, I can basically skip the pep talk.
But I still want to rally against a couple of issues that I read about pretty often, that are related to recruiting and player development.
First, when it comes to recruiting evaluations, way too many college football blog dudez are still way too concerned with what the NFL thinks about college players.
Reasons getting your school's players drafted in the NFL is important to building a good college team:
1. A player's ultimate goal is typically professional football. If you can show playing for your program has a correlation with players going to the NFL, talented recruits will be more likely to sign with your program
2. Those players can further provide visibility and recruiting for your program as they play in the NFL, and hopefully, become stars in the league.
No doubt the NFL can be a useful recruiting tool, particularly to a school without a lot of other built in recruiting advantages. No doubt. But....
Reasons why getting your school's players drafted into the NFL isn't important to building a good college team:
1. You can be a really, really, productive college football player and not have the NFL draft you. Take this roster for example:
RB- Jehuu Caulcrick
WR- Blair White, Mark Dell
TE- Charlie Gantt, Brian Linthicum
OL- Joel Foreman, D.J. Young, Joel Nitchman, Jesse Miller, Roland Martin
DL- Jonal Saint-Dic, Trevor Anderson, Kevin Pickelman, Justin Kershaw
LB- Eric Gordon,
DB- Otis Wiley, Travis Key, Marcus Hyde
P- Aaron Bates
K- Brett Swenson
Over a span of 5 years, that's 21 players from a single team, out of a team's possible starting 24 (both starting 11's plus the two kicking specialists), that you could fill out of guys who made an All Big Ten Team (guys with legit, productive seasons respective to their positions), who didn't get drafted. It's only a few linebackers and a CB or two from filling up a team's 1st string. I don't know if the Franken-team above (hooked up with some depth level fill-ins) would be good enough to win a Big Ten Title in any given year, but they'd absolutely be in the hunt. And, strictly as a college football fan, isn't that much more important to your team's success than the NFL seal of approval?
So when people point to links like this one, I'm always sort of unimpressed because recruiting sites, like you'd probably guess, are pretty good at finding elite athletes who test well in the various metrics at camps and assign them high rankings, and the NFL is also pretty good at finding those same elite athletes who test well and assign them high draft grades. That is, 5 stars are nearly always the freakiest of physical freaks in their classes, and as it turns out, those skills don't typically atrophy over the 3 or 4 years they're in college, even if their production doesn't match up to what you'd expect. This is how, as an extreme example, Bryce Brown gets drafted.
And Brown is another 5 star talent who found his way into the NFL draft, even if it was only in the 7th round. This anecdote would fuel the narrative that high star rankings will get a player to the League, but as any Volunteer or Wildcat fan would tell you, that didn't especially help Tennessee or Kansas State on the field. Again, anecdotal evidence to be sure, but particularly compelling anecdotal evidence in this case.
I've always found this methodology of rating recruiting through the NFL lens to be pretty nonsensical, yet popular among lots of football scribes. So that's the first thing I'm trying to challenge with this look at the recruiting/development system.
The second thing I'm trying to challenge is arithmetic. And math has a mean right hook.
This argument is provided clearly and prominently over the years by Matt Hinton here, and Brian Cook here. The basic argument is look, if you take any given team, it's highly likely that the team will have a higher ratio of successful 2 and 3 star players than 4 and 5 star players, but that doesn't necessarily mean anything.
For example look at MSU:
|Star rankings for Michigan State's All Big Ten players 2007-2011
That's only one consensus 5-star player who made All-Big Ten, compared to 47 players at 3-stars or under! Case closed! But... not quite. See, while they've only had that one consensus five star player, MSU has brought in a bunch of 3 stars over the years (probably something between 60-90 recruits depending on how far back you go ). So when it comes to 5-stars MSU is 1 for 1 in developing ABT (100% of players in an admitted sample size of one), when it comes to turning 3-stars into ABT, it's only like, 19 for 80 (about 24% of players).
Hinton drops this chart that makes a similar point:
Arguing against this chart is made tougher when I don't have proper data to respond on hand (he's using All-American players, I'm using All-Big Ten players. These apples and oranges will not compare well.) But the point is a good one to keep in mind when discussing recruiting: there are many, many, more 2 and 3 star recruits than 4 and 5 star recruits.
For now, I would just say this: While the odds get much lower the further you go down, that even the most certain bets of the independent recruiting prognosticators, 5-star players, yields only an 8% hit rate, is not much too brag about.
And if we accept Hinton's categorization and:
...consider the initial grade as a kind of investment — a projection of the how likely a player is of becoming an elite contributor compared to rest of the field — well, you'd put your money with the "experts" over the chances of finding the proverbial diamond in the rough every time...
There is still room to argue that this articulation misses the point. Regardless of these site's opinions, a good coaching staff can find ways, through scheme, through training, through scouting, and yes, through luck, of beating those long odds, even consistently beating those long odds.Recruiting is not poker or blackjack, where math can be fairly basic and constant.
A bad coaching staff, through deficiencies in those same areas, can waste away even the relatively heightened odds of great recruiting classes to bad performances. I'd argue in figuring out whether a player is likely to turn out, it's much more valuable to first evaluate the coaching staff he will be playing for, then to look to recruiting experts. It turns out coaches can be good and bad investors as well (*NO JOHNELLE BANKRUPTCY JOKES PLEASE OK MAYBE SOME IF YOU FEEL LIKE IT*). A good coaching staff with a 3 star will probably get you further than a bad coaching staff with a 4 star.
Accepting the relatively low predictive power of recruiting websites and putting your forecasting trust or distrust in your coach or school's empirical results might be the best bet in the Big Ten.
So now, with updated and expanded data, I'm again analyzing recruiting services and Big Ten coaching staffs (except uh, one of them) with the help of the conference's All Big Ten awards.
Just to remind everyone, here's the methodology for this from last year's post:
All Big Ten Conference Selections (1st team, 2nd team, honorable mention) players are included if they make either the coach's or media lists. If they make both lists their higher selection is listed (a player who is on the coach's 1st team and a media's 2nd team is listed as a 1st team in my data).
This is a better sample of production and development than say, NFL draft status or all American lists because it is based on college production, is from people who watch most of, if not all, the B1G teams, and it provides a large enough sample size to look at, where even programs like Indiana or Michigan can get enough representation.
It also is a relatively good list of a team's best performers in relation to the league. It gets the best handful at their position, but doesn't include too many players. For example, of the 2010 MSU team, the most egregious 'snubs' were probably Keshawn Martin and Le'veon Bell. Did either of them have that good of seasons at their positions? I don't think so, they were average-to-good skill position backups last year. You could have thrown them on honorable mention, but they certainly weren't more important to the Spartan's success than Cousins, Dell, Cunningham, Foreman, or Worthy, the guys who actually did make honorable mention.
The 1st, 2nd, honorable mention rankings also provides a decent 'great starter-very good starter-good starter' framework. It also builds in a elasticity for the strength of a position across the conference (for example, Kirk Cousins was probably better than just a 'good' starter last year, but the strength at the QB position across the big ten moved him down the rankings, as it should have).
As to whether we can trust coaches who may or may not be filling out these ballots themselves or members of the media to make the right choices, I don't know, all I can say is that it seems pretty accurate to me. If someone wants to deliver a devastating critique of the all conference awards as biased or untrue, to prove me wrong, go for it.
What's up with the half stars:
Recruiting star rankings were taken from Scout and Rivals. Both of a player's stars were taken, added, then divided by two. Obviously in cases where the two sites disagreed (rivals gives a 2 star, scout gives a 3), you get a 2.5* 'consensus ranking'.
This was the same methodology used in calculating the class rankings (adding scout and rivals, then dividing).
Human error is inevitable. As a result, is it possible I might have made a mistake in awarding some player a higher or lower star rating or missed a player's inclusion in the all big ten sections despite double checking? Sure is. Certain walk-on or transfer players were especially likely to create errors. I don't think any possible errors are game-breaking to my conclusions. If you, brave reader, find a mistake ("No way Michael Hoomanawanui is only a 2.5*!" I agree, that's a 5* name), just comment and I'll fix it. Thanks.
With some updates for this year's post:
The Mark Dantonio era. This means awards from 2007-2011, but includes recruiting data from way back in 2003 to make sure that the players of Dantonio's first teams are included. Yes, this means were heading back to the early days of Johnelle and, yes, his recruiting classes are just as badly thought of as you remembered them to be.
Everyone in the Big Ten except Nebraska. Sorry Nebraskbros, unlike the Big Ten Network, I just feel weird annexing your time in the Big 12 into this stuff, and don't feel like your one year in the B1G is really a large enough sample size. Maybe next year. Or something.
There was some debate about this in the comments of the 1st post. To refresh, if a player moves up the All Big Ten teams from year to year (say from 3rd team in 2007 to 1st team in 2008) they get 'double counted' because I think getting that player up the productivity ranking should be rewarded. On the flip side, if a player gets second team in 2007 and then is named 2nd team or 3rd team the following year, they are only 'single counted' because maintaining a player's level, or having them regress from year to year, isn't something I want these rankings to reward as much. There are valid arguments to just count everyone every time that they make a list, but I'm stubborn on this, so DEAL WITH IT INTERNET.
Da- Ok. But seriously, let's get going.
****If you want the raw player data (the long All Big Ten lists of players, and positions, and star ratings) go to this fanpost here.***
Everything else, stay where you are.
1. Conference Recruiting Rankings
|Conference Recruiting Rankings 2003-2012|
A common criticism of Michigan State is that they must increase their recruiting prowess to find continued success. This usually misses two points: 1. How much recruiting has already picked up (in these site's eyes), both from 2003 and from 2007 and 2. How well MSU, UI, and UW have done without 'taking the next step' in recruiting.
2. Raw Player Development Rankings (2007-2011)
|Rankings||Teams||1st team||2nd team||3rd team||Total||Avg per year|
One big lesson from last year's post still stands: there's some value in these website's recruiting rankings, for example, the top seven teams in average recruiting rank over the past six years are also the top seven teams in these player development rankings. But! But, but, but. They don't follow the distribution you'd be led to believe they would. In fact, it's almost the opposite of expectations if you were to look at recruiting rankings. Similar conclusions can be drawn from the conference's bottom four teams.
Let's use these two charts to grade out the Big Ten teams.
Wisconsin: The best. Highest performance at the top (most 1st team players), best depth (most players overall) and doing it all with what Rivals and Scout think are the 4th best classes out of these 11 teams. Between this performance and their on-field success there's little reason to believe they will be seeing much drop off. Yes, they lost some significant staff, but so has nearly everyone at the top of these rankings. Predict their drop from the top of the Big Ten at your own risk.
Iowa: The overachiever. A lack of what recruiting sites see as top end talent has lead to a relative shortage of 1st team ABT players compared to most other top teams high in these rankings. But they make up for this by being incredible deep in the lower tiers of ABT productivity (the most 2nd +3rd team players in the conference). They shook up their coordinators, but as we'll see later, they may have had good reason to.
Michigan State: The high riser. MSU has, year by year, been passing teams on their way up these rankings. Up til recently, they relied heavily on a massive crop of second and third team players to find Big Ten success, but a strong push of 1st team players last year could mean bigger things for the future now that JLS recruiting classes have graduated, leaving MD's recruiting classes in their place.
Penn State: The consistent performer. At its worst over the years, Penn State still put out at least average contributions to these ABT lists, and were probably the closest team to matching up with what recruiting sites would have predicted for them. At their best, they had the most successful year in the time-frame of this metric with 21 players making an ABT list in 2008. But this might have lead to a deceiving spot in this ranking, not matched by their recent success.
Ohio State: The elitist. They will turn out a lot of high level talent. They will use that high level talent to win a lot of games. But they won't produce the depth of talent quite like you'd expect the top recruiting team in the conference would. This can cause problems when that elite talent is injured (or suspended).
Michigan: The underachiever. No one does less with more than Michigan. Even their best teams have lacked either depth (in 2007) or top end talent (in 2011). The three Rich Rod years were dog crap. If you're an optimist, you see that Hoke won 11 games with only 12 ABT performers. If you're a pessimist, you see that Hoke only had 12 ABT performers on an 11 win team. Which of those views is right will likely be found out over the next year or two.
Illinois: The Ron Zook. He did better than you thought he did. He basically performed at an average level over a five year period. He performed about how he should have been expected to perform given recruiting predictions, and above Illinois football's recent history. And yet, it was still totally clear that he needed to be fired. Ron Zook.
Purdue:The indie production. No real star power, only an average amount of second level talent, but bolstered by a lot of small bit parts.They've outperformed their expected results, but not by much. It's ok, you've probably never heard of them.
Northwestern: The Huh. Recruiting sites don't think they recruit very well. This ranking doesn't think they develop players very well. Their scheme or in-game coaching doesn't seem to be anything special. This team should suck. And yet, they've averaged a little over 7 wins a season over the last five years. Huh. *shrugs, throws hands up, blows confused raspberry*
Minnesota: The 'Gophy, no!' Here's a question: someone asks you to make a 'top ten Golden Gophers football players of the last half decade' list. How many names can you write down before you scream out, "I just can't lie to myself anymore!" and concede that Minnesota has brought almost no one worth the price of admission into the Big Ten in the last five years? I'd probably write down Eric Decker, and then write down Marqueis Gray before crossing his name back out because Gray hasn't actually really done anything that special yet (him going bonkers against MSU being very much the exception and not the rule). And... that's about it. Minnesota fever! Catch it!
Indiana: LOL, no one cares about Hoosier football.
How can we closer look at MSU with this data?
The most important teams, comparatively speaking, for MSU in this study are Wisconsin, Iowa, Penn State, Ohio State, and Michigan.
Dantonio talks a lot about trying to emulate Iowa and Wisconsin as models for MSU football. By the end of this piece, it should be even more clear why he emphasizes them.
One question to ask is, "how reliant are you on the upper tier of your recruiting classes?". An easy way to look at this is take a school's total stars from their ABT players and divide by the number of ABT players, then, compare with their average recruiting site stars (I went back to '03 on the recruiting classes).
|Top 6 In-Depth Stars Breakdown|
|Teams||1st Team Average Stars:||2nd Team Average Stars:||Hon. Mention Team Average Stars:||Average ABT stars||Average Class stars||Class-ABT|
I'm left with a couple of conclusions.
1. New coaching staffs for both programs could lead to fans of both fan bases to be dismissive of all this as just history, but it should probably concern UM and OSU fans, at least a little, that to place a player on an ABT list, on average, that player had to outpace their recent recruiting averages by about .15 a star, in stark contrast with the rest of the conference's heavies. Meaning, if you aren't one of the highlights of their classes, it's less likely that you'll make an ABT team.
2. On the other side, with the exception of Iowa whose ABT players are right in line with the rating of their entire recruiting classes, the other three schools, to notable amounts, show much less reliance on the 'best' recruits in their classes, even to the possible bias towards some of the lower recruits in their classes.
3. At a certain point, observations from pundits that no one recruits in the same tier of Michigan and Ohio State (and to a lesser extent, Penn State) just becomes so much nonsense. Yes, Wisconsin, Iowa, and Michigan State haven't recently, and maybe never will stack up with UM and OSU in top class rankings on these recruiting websites. But at least with the current staffs, over the last half decade to a decade, it just hasn't really mattered.
As we can see, over the past decade WIS, IOWA, and MSU have recruited about a rate of .75 stars less per class than UM and OSU yet they've put more (many more, really) players on these All Big Ten teams who, on average, are about 1 star below their OSU and UM counterparts. In other words, contra to the reports of recruiting websites there's a really strong argument that the average standout MSU, WIS, or IOWA player is worth about a star more than the average standout UM or OSU player (PSU can make the same claim at lower rates, their players are about .3 less stars per class, but are about .7 stars better per ABT player). If this would lead you to believe that these 6 teams should be considered to be on at least the same tier competitively right now, regardless of what Urban Meyer and Brady Hoke have done in recruiting, well... yeah.
Who's on their way up, who's on their way down?
I wish I had a 6th year of data from Dantonio to give me an even 3-3 split on something like this, but, 'til next year I guess (I could go back to 2006, but adding in JLS' stellar 4 ABT selections would just make MD look even better in such a split up, and hasn't that comparison been unflattering enough to Johnelle already?). But we can look at this data in two eras and see which teams have maybe built reputations back in '07 and '08 which might not be relevant now, and which teams have done well closer to the present which might be more relevant. I choose to split the years 2-3 instead of 3-2 because... man, i just flipped a coin. There are a handful of teams who get heavily affected by this distribution, I'll talk about them.
|Teams||2007||2008||2009||2010||2011||first two years||last three years||Rising/Falling|
Michigan State: Hey hey, look who it is at the top of the list! Obviously, the last two years are carrying a ton of weight here, but I'm almost as impressed by the relatively depth of this team during their 'basically .500' years (7-6/6-7). 19 All Big Ten players might be where this program tops out for now (and that's no put-down, that's a ton of players), but that is the type of year by year breakdown that builds a year-in, year-out contender. Hard to ask for much more from this staff up til this point.
Wisconsin: Still really, really, good. The most players in the conference on average over the past three years. I'm going keep saying this until they have a bad year: this team is the gold standard in the conference right now. Consistent. Talented. Consistently talented.
Iowa: Still better at turning out talent than their recent records would lead you to believe (that 2010 squad is probably on the all time underachievers list). Be careful with this team though, as you can see they're on a downward slope from that very strong 2009. This is one of the schools who flips from 'riser' to 'faller' if you flip around the distribution of years. You should respect their ability to develop talent and create a dangerous team even if in-game coaching mistakes will likely lose them a few games a year (or, at least has in the past)
Purdue: The other team who goes from Riser to Faller if you switch from 2-3 to 3-2. The difference between this team and Iowa is, though they are a riser, they're rising from such a low baseline (8 players a year to 9.67 players a year) that I just can't get to effusive. The drop from year to year over the last three years is another red flag. Lots of people seem to be buying this team as a challenger in the Leaders, even with 1 (or 2) teams ineligible, I'm not sure I buy it. Credit where credit is due for overall player development progress but lots to be wary about as well.
Illinois: The Ronald Zook Difference! I said earlier that Zook gets a bit of an unfair rap, and admittedly this doesn't do much to strengthen my argument. If Purdue went from ok to pretty good, then Illinois did the opposite over the last three years. The awful 2009 year is an anchor around his neck in this overview, but, well, sometimes you get punished for only putting only 5 players onto 6 separate All Big Ten lists. With that said, 8-10 players a year seems like a pretty good number for a school like Illinois. I'm curious to see if Coach Beck can exceed that.
Minnesota: Oh Minnesota, literally from bad to worse. I hope that program catches a break one of these days, even if this does bring into stark relief just how dumb the firing of Glen Mason was. Kill's had a lot of success with rebuild jobs, but make no mistake this is a REBUILD JOB. He'll need to do better than a horrible 2 ABT players to see much success in the Big Ten.
Michigan: All things considered, this analysis has probably been harder on Michigan than anyone else, but I have good news Wolverine fans! Despite your placement on this list as a 'faller' you've had year by year improvement from a rock-bottom 4 ABT players in 2009 to a sizable 12 in 2011. You still aren't that close to the heights of the Lloyd Carr days, but you certainly weren't going to get there with Rich Rod. For a team whose closest Big Ten comparison over the past half a decade is probably like, Illinois or Northwestern, another strong year would get UM up out of the middle of the Big Ten pack. I, like many, will be watching Brady Hoke's second season with great interest.
Penn State: Uh, whatever I have to say about this team could be irrelevant by tomorrow morning. That said, they're being held up mostly by a conference best 21 players in 2008 and have been on a downward slope even not considering all the off the field turmoil. If you do bring the off the field stuff into consideration... uh, I'mma just climb into this concrete bunker now- Bye guys!
In summation, I think there are a couple of important points followers of the Big Ten conference can draw from this:
A. There does seem to be a minimum of required 'cred' you have to have with the recruiting sites to succeed in the Big Ten but this line is much closer to an average of 3 stars (maybe even an average of 2.75 stars) than it is to 3.3, 3.5, 3.7 stars. Trying to use recruiting rankings to get you insight into who should be in the Big Ten's top tier vs. its middle tier has been fool's gold in recent years.
B. Though most teams have, at some level or another, undergone coaching changes in the past few years, any consideration of recruiting standings would do very well to toss considerable extra credit the way of Iowa, Michigan State and Wisconsin, and likewise would probably be prudent to take some shine off the efforts of Ohio State and Michigan until proven otherwise.
C. High selections of all big ten players on a team doesn't guarantee success, but it will at least get you mediocrity, and low levels will almost certainly guarantee failure.
D. Predictive power from empirical evidence is usually a tricky proposition. With that said 'data > no data', and I'm happy to let the next few seasons prove this post right or wrong.
Hope you enjoyed this!