or How I stopped Worrying and Learned How to Love the NFL Draft ^[1]

[ED: Craig Ross submitted this a month ago and I finally got around to publishing it]

Certainly, as Brian Cook has argued (and everyone seems to agree), high school five-star players are more likely to be successful than four-star players and 4-star guys are likely to fare better than those with three stars or fewer. As a broad brush, this is easily provable. The problem, of course, is that there are not a lot of 5-star guys (around 25 to a max of 30 in any year) and four-star guys comprise the next 250 or so candidates in any individual year. Let’s call it 300 kids who (more or less) 50 teams are after. Where OSU, Bama, LSU, Clemson and Georgia are picking off a lot of these players, that does not leave a lot for other schools.

Start with the 2020 recruiting year as a way of evaluating the problem for the average to very good Power Five team.

So, these 5 schools received LOIs 19 of 30 five stars in the year, with Texas AM getting two, Oregon 3, and six other schools getting one. Of the 130 D-1 schools, only 7 convinced more than one five-star to attend.

This bunching of talent continues with four- star recruits, though with so many more such players the dispersion is broader. Still, the five schools listed snapped up 72 of the 4-star guys. That’s about a quarter of the pool and by the time you scroll down to number 25 on the overall team rankings, these four- star guys become thin in number. Iowa, at number 35, has one four-star recruit. Minnesota, coming off a successful season, secured two. Northwestern got one, MSU none, Indiana one, Purdue, at # 33 none, Illinois, none. Historically successful UCLA in a talent rich area brought in no five-star guys, and a mere 4 players listed as four stars.

Some of this might be chalked up to the decay of the rust belt. Populations have moved out of Flint, Michigan and Lorain, Ohio, to go south. Some of it might relate to Moms and CTE. Certainly, in Michigan, the number of participants in high school football is not what it was in the past. In 2007 there were 46,000+ boys playing high school football in the state. In 2018, it was down to 33,500. For Michigan and MSU, certainly, pulling a kid from Georgia is going to be a lot more difficult than pulling a kid from Michigan. The Michigan kid grew up a fan of the Big Ten, probably followed the in-state schools, and is used to weather. Distance from family is no problem.

[After THE JUMP: Factors]

Distance from Talent

The latter condition should be underscored since an SI study from 2009 showed that of the nine schools that won 50 or more games in the 2004-08 period, seven “signed more than half their recruits…from within their home state or within 200 miles from campus” and the eighth signed 49% of their class from such a radius.

Looking at it from another way, the same author (Andy Staples) cites a study from doctoral candidates from FSU who concluded that most important factor in choice of schools for recruits was “distance from home.” The academic reputation of the school provided a “miniscule bump.” Surprisingly, depth chart considerations and the number of players a school sent to the NFL were not primary considerations. Nor was the recent performance of the school. I am a bit skeptical, but I do buy into the “distance from home” part of the equation.

This is a major problem for UM and MSU who tend to share loyalties in the state. Ohio, a state with (on average) twice as many quality players as Michigan in any individual year, has also declined in student-football participation, though it is around 40,000 as far as I can tell. But Ohio has the state to itself---and more or less owns the hearts and minds of its denizens. Even assuming MSU “owns the hearts” (though never the minds) of a mere 25% of the state, OSU starts with a home-grown pool of 40,000. Michigan? Around 25,000. [Growing up in Ohio, all of my siblings despised the Buckeyes. They were, more or less, Mussolini. But we were outliers.]

It isn’t random that Jim Harbaugh’s attempt to run camps all over the country, and particularly the south, went afoul. While complaints from (primarily southern) coaches were generally couched in having to work/compete during their (relative) down times, the truth is that the SEC just didn’t want schools from the north to be able to compete. These camps were, incidentally, the best thing that could happen for kids interested in playing the game since they were inexpensive (as little as $ 10), allowed access to top-flight coaching, and allowed borderline high school players access to smaller programs, where Division 2 and 3 coaches could both help and evaluate the talent at the camps. The NCAA, not really caring about the players (per usual), put an end to it, regardless of the value to kids.

Scott Bell recently wrote a long article about football recruiting that I highly recommend. His article contends, essentially, that Michigan football recruiting has been pretty good, better than anyone in the Big Ten (he has a slight lean to UM over PSU) except for OSU. Bell argues that the fact that OSU has been killing it on the recruiting trail, should not minimize the reality that Michigan has fared well. His analysis shows:

1. OSU’s average class in the past five years in the Big Ten, is 1.2; that is, they have had the most highly regarded class 4 of five years. Michigan is second at 2.0 (one first, three seconds and a third) and PSU is third at 2.8.
2. Using a star-per-recruit ranking, OSU was #1 in all of the Harbaugh years, UM was 2.4, PSU at 2.8. Nebraska was still fourth (4.5) and Wisconsin passed MSU to be fifth at 5.8.

Bell then argues it makes no sense to compare Michigan to programs in the south because it is a “waste of time.”

Northern programs have some built-in disadvantages when it comes to recruiting—geography, weather, lack of nearby elite fertile talent pipelines, etc. Northern programs with academic standards have even more built-in disadvantages. What do Joe Burrow and Justin Fields have in common, outside of being really, really good at football? Their final exams consist of completing the first three levels of Mavis Beacon Teaches Typing and sending in 250-word essays on James And The Giant Peach to their Transcendent Literature Of The 20th Century 401 professors. And hey, I’m not knocking their priorities. If I’m a five-star athlete that knows my future is in athletics, I’d have a hard time resisting the temptation to do the exact same thing. But just keep things like that in mind when you’re coming up with what you think reasonable expectations should be for Michigan’s recruiting.

How's Recruiting Been?

I might go out on a very short limb and add “cash” to the list of incentives but Bell seems less jaded than me or just sticking to the known knowns, as opposed to the known unknowns. Despite this, Bell has compiled a listing of NFL draftees in the 2016-2020 period. Bama is number one with 48 total, OSU has 44, LSU has 37, Florida has 32, then Clemson and Michigan have 31. PSU comes in 11^th with 23.

Michigan’s Big Ten record over the interim is better than most fans surmise. Michigan has been 32-12, same as PSU, and better than everyone except OSU and Wisconsin, not even adjusting for the absurd losses Michigan had in the interim—PuntGate, Absurdly-Taken-At- Iowa-By John O’Neil-TD Gate, SpotGate, etc. As to Wisconsin, he writes:

Wisconsin has only had to play 14 regular season games against Big Ten East foes over the past five seasons. Michigan has had to play more than double that mark. Wisconsin is 9-5 in those crossover games, compared to 25-5 in games within its own division (Wisconsin is 7-0 against Michigan State, Maryland, Indiana and Rutgers. It is 2-5 against Michigan, Penn State, Ohio State.) I didn’t include Big Ten championship games in the above numbers so it would be an even 34-game sample for everyone. But if I did, Wisconsin’s record against Big Ten East teams would actually be bumped down to 9-8, with a 2-8 mark against The Big Three. So yeah, let’s just save ourselves all some time and not play the Wisconsin card.

Bell concludes that Michigan football is in pretty good shape—top 12 or so shape—but just isn’t where OSU is at, and the juxtaposition seems sharp to most all of us. Well, all of us. For Bell the answer seems to be “develop players better,” though he gives high marks to Ed Warinner and Michael Zordich on that front and he has high hopes for newly hired Bob Shoop.

I agree that if teams are throwing darts at a dart board the five-star guys (as a group) are better than four stars and those are better than three stars. The algorithm is a good one. Nevertheless, I have long been a skeptic of just how good these evaluations are. As I noted in my second book, The Search for the unified Field Theory, an assistant coach at South Carolina (Brad Lawing) claimed that he could move an average plus high school kid to a four star prospect pretty much whenever he wanted---by whispering the right words in the right places. Plus, everyone who follows recruiting has watched some unknown guy commit to Michigan or Clemson and automatically jump into three star range. It is plain as can be that much of defining recruits is backwards evaluative. Kid is being offered by Michigan and Penn State? Gottta be at least a 3 star. Only offered by Vandy and Pitt? Three star, max.

Lawing claimed “I can take a three-star guy and make him a four-star and I can take a four-star guy and make him a three or two-star with the contacts I have. That’s how ridiculous recruiting on the Internet is. I took Chris Culliver (South Carolina safety) from a three-star to a five-star in three weeks. All you’ve got to do is to talk to the right people.”

Now, there is some sense to concluding a kid committing to OSU after being offered by UM, Bama and USC, is probably damned good, even if you haven’t watched one second of the kid play. The coaches at these schools aren’t dopes and aren’t successful for no reason; and they sure as hell, as a group, should be able to evaluate talent. They are not in the business of bringing in kids who they like, regardless of football ability. But doesn’t this mean that at least some part of the evaluative process is just looking at the success of programs over the past few years and deciding that those classes are the best? Isn’t there, at least in some part, the conclusion that Bama or OSU has the best class next year because they had the best teams last year? And the year before that? Isn’t some part of the equation, at least, “kid is going to OSU---push him up; kid is going to Oregon State---he can’t be that good,” right?

Let’s look at the 2020 draft for a moment. I tracked the first 200 picked---guys who one might think would be assets to any college team---and I then looked at each player’s high school rating using the 247 composite. [I don’t claim this is perfect; there may be one or two transcription errors; but it is very close if not exactly accurate.]

• 15 five-stars
• 31 top-100 four-stars
• 18 top-200 four-stars
• 15 ranked 201-350
• 13 ranked 351-500
• 16 ranked 507-700
• 17 ranked 701-1000
• 40 ranked 1001-2000
• 35 ranked 2001+

So, a 32% hit rate, that is, 32% of the players rated in the top 200 out of high school found their way into the top 200 of the 2020 draft. I think that’s pretty good and not a likely anomaly, though I will check other years at some point. Then, 30.5% of the draft comes from numbers 201-1000, with reasonably homogeneous distribution over these levels. That is, 7.5% of those drafted fell into numbers 201-350; 8.5 % came from a range twice as broad, 701-1000.

On the one hand, while the 32% hit rate is pretty good, the opportunities for finding guys who can really play outside of the four-star range is also pretty good. If one defines a four or five star as a top 350 guy, this means 121 of the top 200 players (at least as defined by the NFL draft) were three star guys or less, guys that UM sure as hell would have a good possibility to recruit. Let’s take this team with draft number in parenthesis followed by the composite number.

Offense

• QB Justin Herbert (#6) #659 or Jordan Love (# 26) # 1,665
• RB Edwards-Helare (# 32) #378
• TE Joshua DeGuera (# 94) #2,163
• OT M. Becton (# 11) # 405, Ezra Cleveland (# 58) #1,812
• OG Robert Hunt (# 39) # 3,212, Jonah Jackson (# 72) # 1,733
• C Hennessey (# 78) #2,228
• WR Justin Jefferson (# 22) # 2,164, Denzel Mims (# 59) # 410

Defense

• DT Devon Hamilton (#73) # 998, L. Fotu (# 114) #575
• DE J. Zuninga (# 79) #636, D J Wonnum (#117) #1,304
• LB I. Simmons (# 8) # 451, K. Murray (# 23) # 434, J. Brooks (# 27) #883
• CB D. Arnette (#19) # 665, J. Gladney (#31) # 1558
• S K. Dugger (# 37) Unrated, A. Winfield (#45) #1,415

If Harbaugh had recruited this class Michigan fans would have gone berserk. Not a single five star. Not a single four star!!! 5>4>3. Hell, some of these guys are probably one stars. This looks like the team assembled by the Cleveland Indians in Major League. Yet, all of them went in the first four rounds of the NFL draft. Eight of the eleven listed offensive guys went in the first two rounds. Seven of the eleven defensive guys went in the first two rounds. Assuming the NFL has any idea of what they are doing, you can build a pretty nice team from guys that were strictly “meh” when being recruited.

I noticed some “crowing” by recruiting analysts post draft, some (if tepid) back patting that they had really nailed it with their high school evaluations. I found this a bit odd, since I would have thought this 32% efficiency hit (for a lack of better words) would be within the realm of expectation. Did this mean, perhaps, that the 2020 draft was an outlier?

Well, I don’t really know, and these things take time; it takes me, at least, a couple of hours to look at a draft list and see where the players were rated when coming out of high school. But I did go back and look at 2019, to see if it looked like 2020. Here’s the top 100 in 2020.

In the second 100 it started to water, but the “crowing” seems justified. 52% of the first round comprised 4 or 5 star guys. How about 2019?

Or, looking to bracketed orderings:

Here the “efficiency” is less obvious. Plainly, more work needs to be done.

My first thought is that even though 5-star guys are better than 4-star guys, etc., in a macro, dart throwing contest, that really doesn’t define the problem in an adequate way. As a fan, I don’t care whether or not, as a group, five stars are more likely to succeed. I care that the guys Michigan brings into the program succeed, at last at some reasonable rate. The truth is there are profound market inefficiencies. There are numerous guys, every year, who are undervalued, some well undervalued. In the 2020 draft Michigan found a couple of them. Josh Uche, # 707, was drafted (# 60) late in the second round. Khaleke Hudson (# 386) went in the fifth round (# 162). Jon Runyan (#1,138) was drafted in round six (#192). The banality is plain. As a Michigan fan I would prefer one to three-star guys who succeed, as opposed to four or five-star guys who don’t. Here’s Michigan’s 2016-2018 recruiting, clumped by their 247 composite ratings. I apologize if I have missed one or two.

1-50

• Rashawn Gary (#1). Drafted # 12. Gary was good, but not as good as we hoped.
• DPJ (# 12). Good but disappointing. Drafted # 187.
• A. Solomon (# 23) Who knows? Transferred to Georgia.
• Bredeson (# 39) Drafted # 143. Very good player, slightly below his fanfare.
• Ruiz (# 43) Drafted # 26 and easily lived up to expectations or better.

Comment: Overall this is a good group, but just good relative to the rankings. Only Ruiz truly lived up to his billing. But all four guys UM kept made game-time contributions. My surmise is that Gary was drafted as much on potential as performance. Guards get devalued in the draft, so maybe Bredson was close to spot-on.

51-100

• L. Villain (# 57) The jury is out.
• B. Peters (#61) Transferred to Illinois.
• David Long (# 65) Drafted # 79. Yep.
• Asiasi (# 77) Transferred to UCLA. Drafted # 91
• D. Singleton (# 79). Transferred to Rutgers
• Onwenu (#87) Drafted # 182. See Bredeson
• Ambry Thomas (# 90). Yep

Overall, pretty good. At least four of these guys will either play or have a chance to play in the NFL. Asiasi, we barely knew ye. Singleton didn’t work (injuries) and Villain is unclear, right now. It seems likely, from an NFL-draft perspective, all of these guys slightly “underachieved.” OTOH at least 3 were valuable assets for UM, the coaches appropriately targeted Asiasi, and Villain is still in play. Peters? You tell me.

101-150

• J. Anthony (# 104) Transfer.
• Ka. Walker (#107) Late faller who transferred to Miss. State
• Filiaga (# 112) Could this be the year?
• A. Hutchinson (# 112). Yep. A definite hit.
• T. Black (#116). Injured. Kept waiting for breakthrough. Transferred to Texas
• McGrone (#118) Yes
• K. Crawford (#118). Not much in AA. Transferred to Cal.
• McCaffery (# 123) Let’s hope.
• La. Hill (# 132). I don’t care what the pros think. He was good at UM.
• Nico Collins (# 138). Yep

All over the place. Four of ten definite hits. Three definite misses. Two still plausible. I think I would guess a 50/50 hit rate in this range and that’s what it looks like. Par?

151-200

• M. Muhammad # 168 Transferred to Houston. Not much in AA
• Spider Sims #170. Transferred to Ga Tech. Ditto
• O. Martin #180. Transferred to Iowa. Ditto
• Blowout. Maybe this is a range JH should avoid.

201-250

• J. Milton # 204 Who knows?
• J. Ross # 211 Looks like a starter this year
• J. Hudson # 231 Transferred to Cincy? Nothing in AA
• Irving-Bey # 249 Transferred. Nothing in AA
• J. K. Powell # 250 Transferred. Not much in AA

For a group of guys solidly in the four-star range, this doesn’t seem too good to me. Of the 8 guys in the 150-250 range, this is disappointing. Only two of the 8 have any chance of game-time value in AA.

251-300

• Malone-Hatcher # 268. Medicaled. Torn Achilles in HS.
• Mayfield # 268 Bingo!!!! Starting tackle and looks certain for the NFL
• Samuels # 269 Transferred to New Mexico State. Nothing in AA except legal troubles.
• Jeter #289 Getting late fast, but still in play.
• Ron Johnson # 298. Getting late faster but still in play.

Comment: Continues the trend of # 100 plus. Mayfield a major asset, probably an NFL player. Otherwise, maybe Johnson or Jeter will break out.

301-350

• J. Hall # 301. Transfered to CMU. No progress at UM.
• C. Kemp # 310. Not an NFL guy, but might yet be a factor for UM
• D. Bush # 312. Drafted # 8. Borderline 3/4 star hits big.
• C. Evans # 334. Pretty good at UM with a viable future.
• Hayes # 335. Possible starter this year at OT.
• St. Juste # 338. Mistake by UM? Starter at Minny and played very well.
• Eubanks # 344. Probable starter at TE with future in NFL

Comment: So, this group, high three stars or low four stars, did very well. Six of the seven have started at UM.

St. Juste is a starter and star at Minnesota.

351-400

• Steuber #365 Has started some. Might start in 2021.
• McDoom # 380 Transferred to USF. Some success at UM.
• Geman Green # 382. Still around.
• J. Woods # 383 Transferred to Duke. Played a bit in AA.
• K. Hudson # 386 Started for two years. Drafted # 162

Comment: Pretty good. Hudson and Steuber certainly had/have value on the field. McDoom and Woods played some, at least.

401-500

• Honigford # 411 Competing this year
• Hawkins # 425 Starter in 2021.
• Paea # 458. Still fighting for a spot
• Kwity Paye # 487 Jackpot
• Mbem-Bosse # 491. Non football problems. Left program

Comment: Maybe a 60% hit rate here. 40% minimum.

501-700

• C. Turner # 508 Will compete for PT this year
• Upshaw # 561 Competing for a spot
• Nate Johnson # 590 Flashes at UM. Transferred to Tenn.
• Faustin # 592 Competing
• Van Sumerin #624 May yet find a role
• Welschoff # 645 Competing
• V. Gray # 700 Probable Starter

Comment: Hard to say at this moment how this group pans out.

701-1500

• Uche # 707 Drafted # 60 by NFL
• K. Davis # 731 Transferred pretty quickly to UAB
• Spanellis # 735 Decent player and great student. Grad transfer to Vandy
• Metellus # 739 Starter for two years, drafted # 205
• Barrett # 751 Probable starter at Viper
• Schoonmaker #796 Will play in 2021
• Mason #817 Anyone want to pass on him? Not me.
• McKeon # 854 Played a lot and started in AA; Undrafted FA
• Kurt Taylor # 917 Transferred early
• Dwumfour #947 Potential starter grad transferred to Rutgers
• Haskins # 975 Will carry the ball a lot
• Gill # 997 Played a lot. Started 14 games. Grad transfer to USF
• German Green # 1,225. Still around
• R. Bell # 1,473! Starter, potential NFL player

Comment: This is surprising. Every one of these guys, save for Davis and Taylor (early transfers) and German Green, made some game-time contribution to the program or is quite likely to. All but the transfers and Green have or are likely to start a game at UM. Some, at least Uche, Metellus, Bell, McKeon, Haskins (my opinion) have legit shots in the NFL.

What are the conclusions from all of this? First, of course, the data points are quite limited. One would need a lot more evidence before coming to much of anything concrete. Still, isn’t it a bit surprising that the guys rated over 700 have tended to pan out, while guys in the 150-300 range have not. Let me guess this, as a starting hypothesis---that players between the 150-1000 range are very hard to distinguish and predict. It is, of course, hard to project most any player. Kids are growing. Competition levels are uneven. Some kids (thinking about guys like Kugler or Kalis) may have been so well coached and trained before coming to UM that their ceilings weren’t far away. Other guys (thinking about Uche or Bell) may have been so new to the game, so raw, that their high school techniques and production would blossom in a more sophisticated environment. Here are some working ideas, until they are proven untenable:

1. Jim Harbaugh and his staff have had a difficult time bringing high level (top 100) guys into the program. In the three-year period from 2016 to 2018 these totaled 12. Of these only 8 lasted for any duration. However, of the eight who stayed seven were productive at Michigan, while one (Villain) remains a question. Looking to the years 2019 and 2020, the trend is downward. Michigan inked 3 top 100 guys in 2019 and none in 2020, about as few in any two-year period as I can recall. That said, the 2019 class looks quite good. At least six guys (Dax Hill, Charbonnet, Cornelius Johnson, Giles Jackson, Eric All, Mike Sainristl) showed something on the field, quite a few for a freshman class. Rumors about others are strong including # 554 Zach Carpenter, who may be destined to start at center. It is hard for me to imagine that Chris Hinton or Mazi Smith won’t be productive. I really like the tape of Quintel Kent (#958) and since my last prediction on point (Haskins) worked out, I guess I am willing to go out on another limb where I probably shouldn’t be. What is undeniable (assuming I haven’t miscounted) is that UM, over a five-year period, brought in 15 top 100 guys.

I find this hard to believe but maybe Christian Dawkins is right, that recruiting in football is “even worse” than in basketball, that unless you are willing to provide benefits to players beyond what the NCAA allows you are not getting your foot in the door. Still, we are not talking about 15 five-star guys. We are talking about 15 guys in the top 100, three of whom were 5-stars.

2. Ohio State, as a salient comparison (not counting transfers like Justin Fields) brought in 44 top-100 guys over the same period, and that included 14 in a single year, 2018. The “good news” is that they have signed “only” 11 in the past two years. Is this a gap (44 to 15) that can be overcome? I answer this question with a “yes,” if Michigan is really, really good at evaluating talent and finding the talent that can grow. The “team” I pulled out from the recent NFL draft, for certain, would likely trample OSU, assuming the NFL knows what it is doing.

Plus, what really has been the difference between UM and OSU over the Harbaugh era? Certainly, it is fair to suggest that OSU just has better players. But my guess is the core of the problem has been QB play. OSU has killed it at the position (Joe Burrow, the # 1 pick, was blocked on the depth chart at OSU) while Michigan has struggled over the Harbaugh era pretty much every single year. That takes but one player to change and, who knows, maybe that player is J. J. McCarthy, Joe Milton or Dylan McCaffrey. Or all of them.

3. The actual good news is that Michigan has been better than reasonable in evaluating talent in the # 700+ range, a place where it would not generally be competing against OSU or (say) Alabama. And here’s a sliver of “inside” rumor sort of news, the sort of thing I rarely have. I have been told that Michigan is attempting to develop a data base that will allow them to ascertain what sort of factors predict success and failure. Now, I have no idea what this data base looks like or what any correlations might show. But, as far as I am concerned, this is the sort of thing that should accrue/be going on to have any chance to beat the market. It could be astrological signs or the day of the week the kid was born. It could be how long his arms are or how little he can bench press. It could be looking to a coach with a particularly good record of evaluation. I am not saying that such correlations exist, but in the wake of Bill James and Billy Bean this is an arena that seems plausible. Simplifying, Bean originally targeted guys who “didn’t look like baseball players” who walked a lot and did not have the sort of flash that other teams might value. Many years later, all baseball teams have computer/stats guys looking for ways to beat the market. That Michigan is doing this, or so I believe, is a good sign. Where many NFL teams, according to one insider “doesn’t know what the word ‘analytics’ means,” I was pleased that Jim Harbaugh is at least looking to such an attempt.

Certainly, this is a harder (and maybe not possible) enterprise in football than baseball, a lot harder, where the game lacks linearity and data is often soft to non-existent; where coaches are evaluating high school kids, where coaching and competition is all over the board. For his part, Bean preferred to draft college players, where the vagaries of growth and competition were minimized. That’s a luxury Jim Harbaugh does not have. That said, the attempt is the right attempt, and I believe Michigan is serious in trying to find this particular unified field theory.

Recently, I was sent a scholarly paper from the University of Pennsylvania, titled “The Loser’s Curse: Decision Making and Market Efficiency in the National Football League Draft.” This is a 2013 document from stats guys, William Massey and Richard Thaler. It was a revelation for me, or more accurately, made my head spin. Massey and Thaler looked at NFL drafts over a number of years and concluded, to boil it down, the NFL’s “ability to predict performance is quite low…[and that] the surplus value of draft picks actually increases throughout the first round, i.e., late first round picks generate more value than early first round picks.” Now for the authors idea of “value” is the delta of on-field production and salary so this doesn’t necessarily mean that the absolute “values” of players declined. Still, over a five-year period they point out that first round draftees had more seasons with zero starts (15.3%) than selections to the Pro Bowl (12.8%). Here are some other findings in this paper:

1. “[NFL] scouts predict exceptional performance by college players in the NFL more frequently than is warranted and that among these players to be predicted to be superstars there is no relation between ratings and performance.” Note that these are college kids, older and with a greater homogeneity of competition than high school kids. It would seem if the NFL struggles at this level to predict success, then colleges, with fewer resources and evaluating a less mature group of subjects should really struggle.

2. Massey and Thaler asked this question regarding the NFL draft. “What is the likelihood that a player is better than the next player chosen at his position.” In other words, assume a team has Linebacker A at the top of its list and Linebacker B as the next best alternative. What is the likelihood that Linebacker A is truly a better player than Linebacker B? My instinctive guess to this question would be “about 70%” given the energy that NFL teams have in creating these evaluations. The actual answer is 52%, a flummoxing one for me. They write, “Across all rounds, all positions, all years, the chance that a player proves to be better than the next best alternative is only slightly better than a coin flip. This simple observation suggests a discrepancy between the teams’ perceived and actual ability to discriminate between prospective players.”

I might add to this, that the team/person who has “some” ability to make distinctions is the king of the world. For example, if you are Michigan looking for a wide receiver, why waste your limited time and resources chasing after prospect # 10-- who is likely to go to Bama or Clemson anyway and receive a paycheck--- when prospect # 1,473 (Ronnie Bell) will be a completely acceptable and perhaps superior alternative who no team is looking at? Why waste your time/limited resources on highly rated running back who dumps you for OSU (Mike Webber) when Hassan Haskins (at least as good) is sitting in the wings, with none of the sturm or drang? Of course, everyone could see that Webber had value. No one, apparently, saw that Haskins did. Michigan just needs this king of the world or some way of sorting.

3. Aside from Massey and Thaler’s concept of “value,” in order to save my sanity they do concede/conclude that “teams do have some ability to predict player performance and this performance is related to draft order.” OK, we are back to 5 stars>4 stars>3,2,1 stars. At least we know that much. This is supported by a 538 article from 2016 concluding that “there is evidence teams are getting better talent in earlier rounds.” This conclusion, looking to career longevity and Pro Bowl appearances, is tempered with the idea that players with “good pedigrees and name recognition are being given preferential treatment when awards are granted, in the absence of meaningful performance data (like what exists for offensive skill players) to challenge our perceptions.” Meanwhile the authors go on to point out that undrafted players continue to make a significant mark in the NFL, with these increasing from 497 in 2005 to 746 in 2014. [2]

At this point I am left with questions, not answers. Maybe there are no answers and the best one can do is 5>4>3. However, I did go back to the question where I think analysts at Michigan should go, “How did we miss out on this guy? Why didn’t we pay any attention to him?” To me, this is a much more utile question than “How did we miss out on Joey Bosa?” Or even “Why can’t we recruit Joey Boasa?” In my superficial look to unrated HS guys who went high in the NFL draft, there does seem to be some tendency to “miss” small town guys playing in the middle of nowhere. It is easy to understand how T. J Hockenson (# 8 in NFL draft, #1,468 on the 247 composite) would be overlooked on the national scene, coming from Chariton, Iowa, a town of 4,200. Same for (say) Chris Lindstrom (#14 NFL, # 1,037 on 247) from Dudley, Massachusetts, a town under 12,000. That Iowa and ISU were the only suitors for Hockenson is understandable. That BC was the only offer to Lindstrom is understandable.

But what about a guy like Sean Murphy-Bunting? He was unrated out of high school, yet was drafted # 39 by Tampa Bay and was a rookie All Pro last year. This guy was right under UM’s nose, at a position of perpetual difficulty to fill (CB), playing at Chippewa Valley High Scholl in Macomb Township. CVHS is a largish high school (2,400) in a city of 100k. That area/high school is regularly scouted. Murphy-Bunting had an offer from Ferris State, and nothing from D-1, until CMU came in with a grey-shirt proposal. He was a success in year one and left after year three, a star in the MAC. In the 2019 draft alone, there are numerous, similar stories. Looking to all of these, I have a working theory, but I will pass it along to UM once/if I feel there is any merit to it.

In the end, perhaps the answer is NIL. Hell, MGoCorp can be formed as a GoFundMe with shares being purchased at $ 100 each. For that you get to attend a shareholder’s meeting once every year and vote on a board of directors who will serve at no pay. This board will take the monies raised and communicate with agents of HS players with their offers. Now, it is true that the player has to provide “fair market value.” But where Feng Shui expert can make $ 10k per lecture, or where Coach K commands $ 100k plus per lecture, I think a once or twice per year event at MGoCorp should suffice. I would assume, starting in 2021, all we need to know is 5>4>3.

Jon Rowe says I need some conclusions. Well, I don’t have any, at least yet, except that I think the market has enough slippage or inefficiency to beat. But whether it can be beaten is another matter, in this (sorry) non-conclusion, conclusion. However, this is just part one. I think there may be something to learn from the recruiting results of Nebraska and Wisconsin. The gurus, routinely I believe, have Nebraska recruiting at a higher level. Yet, Wisconsin has been a far superior program. That’s my next stop.

[1] My friend Jon Rowe killed my original title which contained some hoo-haw about “market inefficiencies.” Since I don’t have any firm conclusions, he thought this was a bit much. Yeah, he is right. He also said cut this document in half, among other criticisms. I tried to be more concise but I didn’t do that. I well let Seth Fisher do that.

[2] My friend Mike Golob makes this point, and it seems well plausible. “In the NFL draft, teams often reach at a position of need, especially in the 1st round, particularly QB's, OT's, and pass rushers (or TE's for bozo central). I'm pretty sure it skews the bust/hit rate. In the later rounds, teams generally go BPA so the bust rate might be lower because of that. “