# McBean 2002, 2003 Final and 2004 Preliminary

Submitted by Meeechigan Dan on September 29th, 2009 at 12:34 PM

Note: The spectacular rollercoaster start to the season distracted me from our McBean effort. Back to work. Source Material: Original Post, Definitions, 2002 Class, Problems.

There was a flurry of concern about how our definitions, particularly the proliferation of McBean two-stars when we rarely recruit any Rivals two-stars, necessarily results in an overall decline in performance for any given class. This is a big deal when you recall that the purpose of this project is to allow us to judge whether a staff’s recruiting and talent development exceeds or falls short of expectations. We can say right now that they will almost always fall short of expectations; the question is how much. One day, we may adopt a variation of the proposed formula from wolfman81 to solve this problem (I say variation because I don’t think it goes far enough to compensate for the two-star problem – see the 2003 class RMS below, which narrows the gap as it should for a great class, but not enough…I think somehow weighting five-star players may be the answer):

Lastly, you asked me about my formula. It's really just the Root-Mean-Square. So add up for each player (star rating)^2. Then divide by total number of players. (This is the mean of the squares.) Now take a square root so that the numbers are comparable.

Example: Compare these 2 person classes (2 4 stars, vs. 1 5-star and 1 3 star)

2 **** -> Avg = 4.0, RMS = 4.0
1 ***** + 1 *** -> Avg = 4.0, RMS = 4.123

I'll ask the question this way. Would you prefer a class that is half 3-stars and half 4-stars (remember, I'm talking about McBeans here--so 12 All-Conference players and 12 servicable backups) or a class that is half 2-stars and half 5-stars (so we have 12 All-Americans in a single class and 12 guys who never play)? I know what my answer is (especially if we consistently recruit and develop that kind of talent).

I am rolling out this stat below the averages for your consideration.

In the end, we decided to finish out our McBean rating effort and go from there. With three classes in hand, we will be able to gauge the two-star problem and either wallow in that misery, as UMFootballCrazy wants, or create an algorithm to compensate the relative value of players, as wolfman81 wants. The Team Ranking analysis for the 2003 class demonstrates the two-star problem clearly…

…but we’re going to finish and circle back. I will probably even finish the 2005 class, even though we have active players.

So, mush.

Here are the first two final classes, 2002 and 2003:

For the 2003 class, Kraus was bumped to a four star, which was near unanimous except to UMFootballCrazy, who doesn’t like to ignore the NFL draft in this instance. I have felt bad for SanDiegoWolverine in the past because he keeps passionately arguing for certain guys and not getting his way – in this instance, he wins as both Rivas gets moved to a four-star (you can’t use the NFL draft as a tie-breaker for kickers/punters and he is a multi-year starter) and Richard gets reclassified as N/A.

JimHarbaughScramble is clearly still grappling with extreme Mundypobia, but sorry JHS, I can’t make a drafted DB a two-star no matter how many times you see this running through your mind:

Here is the preliminary 2004 class, which is our last class that we can call complete (the 2005 class has six active players). There are plenty of issues, and most of them seem to result from overrating our favorite players, like Henne and Hart.

There's a lot to debate in the 2004 class.

I don't like N/A as a rating as it changes your sample size from recruiting class to McBean class. There was a reason that these guys didn't see the field. They need to be 2-McBeans in my opinion. I know that this further kills a class rating...but they represent undeveloped talent. Now maybe they transferred and played elsewhere (Gutierrez for example) but they should be rated somehow OR eliminated from the recruiting class entirely (even the Rivals side). But I don't like that solution as we are trying to correlate (Rivals) recruiting ratings with actual college performance.

Some thoughts on 2003:
I really like these ratings. I renew my N/A objection.

As a preface to my next portion, let's remember that as Rivals thinks about recruits nationally, we also have to think about McBeans nationally. A team MVP need not be a 5-star. You wouldn't argue that the best player on NIU should be a 5-star McBean player. We must take care to rate our favorite players without passion if we intend to truly make this exercise worthwhile.

On to 2004 (let the negging begin):
Henne = 4* McBean. He wasn't All-American. He was never an All Big 10 first teamer (2nd team 2006, Hon Mention 2004). Not a borderline case (and his draft status is too low anyway). IF he had been healthy his senior year...well, I'll just say that my story might change a bit. But he just didn't rise to that level of excellence.

Hart = 4* McBean. Played in the "national shadow" of Adrian Peterson and Darren McFadden. 1st team all B10 in 2004 and 2006, 2nd team All B10 in 2007. But he was not All-American. If we admit borderline, his draft status is too low to truly warrant 5-star rating.

Alan Branch is an interesting borderline case. He wasn't All-American though. I will not be as vehement in changing his rating to 4-star. I believe that the reason that they choose to use DL (instead of DT and DE) is to pick up more of the gaudy stat guys rather than the double-team eating man-beasts. (You can make the same argument about RB instead of TB and FB.)

Rogers, Chetham, Allison - I renew my N/A objection.

(Weeeeeeeeeeeeeeeeeeeeee! Look at all my MGoPoints disappear!)

LOL, no, no. No negging for well-reasoned McBean opinions...

By the letter of the definitions, you are absolutely right about Henne and Hart. I put them at 5 stars to stir the pot.

As nor the N/A, you are aware that the denominator changes for the McBean calculation? I assume you are (as the math guy behind proposing the RMS) and I am guessing you don't like the staff getting a break for recruiting a bust, like Dann O'Neill. But what about Antonio Bass?

Should we differentiate between a N/A injury situation like Bass and a N/A bust like O'Neill perhaps?

I just touched on this above but if we recruit a bust and he eats up a scholarship for 4 or 5 years that is a bad thing and should be represented by a 2-star McBean. If that same bust is encouraged to transfer to EMU (or Northwestern if he likes his academics) then that's better than the first scenario.

It will kill our comparison...all classes will so drastically underperform, we won't be able to compare to Rivals. However, the logic is sound.

To me, the real value is to be able to say: "Did you see that the 2011 class was the 8th class in the team rankings, but the McBean class grades out much higher?" or "Mike Jones came in as a three star and left a five star! What a find."

We may be able to still say the second thing, but the first one will be toast. Maybe the first doesn't matter because I don't have any other teams.

I don't like N/A for just anyone. I do think that it is appropriate for a career ending injury (unless there was significant data before the injury occurred) like you might use for Bass. Career ending injuries are rare and random. In this case I would remove the injured player from both the recruiting class and the McBean class. I don't know that we can think of another career ending injury (like Bass) in the past 15 years.

If you recruit a player and he busts, that should be reflected in the final McBean rating. I don't remember the specifics of O'Neill, but you pretend that he does not exist in the program after his recruitment by using N/A.

As far as the "two star problem" goes, it may not be a problem at all. We have set our "definitions" based on what Rivals actually does based on counting. When we say that a 5-star McBean should almost certainly be an All-American, we do this because there are about 30 rivals 5 stars and 25 All-Americans so that the numbers match up. Most of our 4-Stars are All Big Ten. (Again we do this because 25 All BCS conference team players x 6 BCS conferences = 150 players plus a few more gets to a Rivals top 250--4 star--player.) So the question can be asked, do we recruit the same number of 4-stars (or better) that get all conference recognition annually? Perhaps there is some problem with Rivals where they say: Well if Michigan* offered them, they should be rated higher. So we might expect "underachievement" by some standard. At some point in their careers, do you expect Tate AND Denard** to become All-Conference QBs? Probably not. (Although if Denard gets a wildcat-type package role and sees the field in the slot where he can use his speed on the edge, then you might see both on the All Conf team in different positions.) I guess that the bulk of my complaint is that the coaching staff might say: OMG we need (position) this season and recruit 3 players for a single position. They won't ALL play. But they might ALL be 4-star recruits...especially since they were all offered by Michigan. What we should be looking for is some sort of pattern looking for trends. Comparing the raw numbers may be misleading. In specific, I'd look at three things.

(1) Rivals Ranking - Did we recruit talent?
(2) McBean Ranking - Did that talent develop?
(3) Final number of points in the AP or ESPN poll. (Or some other metric that attempts to answer "Did the talent on the team turn into victories on the field?")

Plotting them with respect to time...graduation year for most of the recruiting class. (So the 2002 recruiting class should be plotted with the 2006 polling data.)

* Obviously, this should apply to other high profile programs as well, like USC, OSU, UF, etc.

** Were the both 4 star prospects? I don't keep up that much with recruiting.

Goodness...for some reason these always turn into novels...

I think the idea of a case by case review of each potential N/A (Bass vs. Taylor Hill vs. Dan O'Neill vs. Clayton Richard) is probably wise. There aren't that many. I think that will be a takeaway from this thread, to look at reviewing the N/As.

I am surprised that we are getting so little play on the 2004 ratings. I think the thread is too long. I think I have to come up with mini-McBean posts that ask about specific players and that's it. Maybe on the board.

First, it is really football season! There is part of me that thinks you should set this down until February and pick it up again then. We are excited enough about the future that we don't want to dwell in the past.

Second, this is very deep, content rich material. People may not have thought about it much (and they are thinking about MSU right now). And this hasn't been a regular enough feature that people know immediately what you want.

Lastly, as we all get used to this conversation and figure out what you want, there will be some lessened debate. As we firm up our definitions, we get to make more straightforward arguments.

Also, I like Misopogon's RR idea (use Rivals Rating instead of star ratings...) however more categories = more debate. (Which may not be bad.) But it is more work (ostensibly for you). It would provide more ammunition for an opening entry in February entitled "McBean 2.0"...

Excellent suggestion. I think, in the interim, I will roll out MGoBoard topics for individual McBean ratings.

McBean: David Harris

Have a short recap of the McBean concept and ask for the player to be rated.

I think that might allow me to accumulate a lot of data, since the contributors now are down to about a dozen serious posters.

I'm not sure we will ever see eye to eye. I like how the N/A's are in there now. If a 5-star QB comes here and doesn't want to wait behind another stud so he transfers; we shouldn't get hit double on that by giving him a 2-star McBean. The Rivals rating will already be higher and thus our expectation of that class. In that sense we are already affected by his transfer. In the same sense, if we take a flyer on a 2-star and he doesn't develop we should be glad he transfers so we can fill he scholarship with someone with a better chance at contributing. In this way we slightly benefit from getting an N/A from the McBean rating which I think accurately reflects the slight benefit to the team.

I also think Henne is not a 5-star. The reason why he was even a 2nd round pick is because of his build, height, arm, etc. . . He set all kinds of records because he started 4 years and was a huge part of our 2006 run. However I think most of us would agree that he had the potential to be something greater. And the fact that his talents weren't utilized or developed to their complete potential here should be reflected by a McBean 4 star.

That's an excellent forgotten point on my part about the transfer of top talent, like Mallett. Should Mallett be a two star? Should the staff get credit for him being a four or five star depending how things turn out at Arkansas? The N/A nuetralizes the transfers.

I don't know. This is a difficult question. I guess there are two arguments to be made (applied to Mallett):

1) He would have turned out here just like he did at Arkansas. So if he is awesome there, he should get a 5-bean rating (or whatever his career merits).

2) He didn't develop here for more than a year. So you can rate him on his freshman season and (perhaps) play the game, "If he had stayed here..." Even in this scenario he has to grade out at least as a 3 star as he was not invisible! In fact, he would fit the multi-year starter mold had he stayed last season, so 4 star gets in the discussion too...

Perhaps my biggest opposition to N/A is that it takes a non-standard situation and refuses to look at it at all. It is the easiest solution, but I don't think it is the best and I don't think that it addresses the purpose of the initial question looking at talent evaluation and on-field results.

I'm surprised I'm only coming to this thread now. Excellent diaries, M-Dan!

Kick me if you explained and I missed it: I was wondering, however, why you chose stars rather than Rivals Rating (RR) as your basis for assessing the players.

RR Stars
6.1 5-star
6.0-5.8 4-star
5.7-5.5 3-star
5.4-5.0 2-star
4.9 1-star

The reason is because we are essentially throwing out a TON of good information that Rivals has provided with their rating system.

Rivals has been using the RR system since 2004. And I think it's a lot more telling than stars as to what they thought of each prospect.

The difference between Brandon Graham and Carlos Brown in the RR system was 0.1 points. The difference between Brandon Graham and Cobrani Mixon was 0.3 points, which is the same drop-off as Carlos Brown to Quintin Patilla. However, your star system rates this gap as the same. If Brown reaches the potential that we expected of Patilla, has he justified his projection? No.

In 2005 Rivals rated Antonio Bass, Mario Manningham and Marques Slocum as 6.0 players. They also rated Brandon Harrison, David Moosman and Justin Schifano as 5.8. All of those players were 4-stars under your ratings. But Rivals was already saying that Super Mario would project just under a 1st round NFL draft pick, and that Harrison and Moosman would be just barely above an all-region selection.

It's not easy converting a 12-point system to a 5-point system. So what if we forget about stars and went with grades?

6.1 A+ Franchise Player; 3 or more years starting, 1st Team All-American, Multiple All-Big Ten, Tiebreaker: goes in 1st Round of NFL Draft
6.0 A All-American Candidate, high-major prospect, long-term impact on the team, Tiebreaker: 1st Day NFL Draft prospect
5.9 A- All-around good player, multi-year contributer, All-Conference candidate, Tiebreaker: Drafted 4th-5th round by NFL team
5.8 B+ Good player, starter by 4th year, considered at peak as All-Conference player, Tiebreaker: NFL prospect in late rounds or undrafted FA
5.7 B Solid player, starter by 5th year, considered a "role player." Ability to make an impact in college, would be a better player for a Mid-Major
5.6 B- Contributer, starter due to circumstances, or dependable "depth player." Typical a "3-star" performer, mid-major starter
5.5 C+ Contributer, but with obvious holes. Would be a middling mid-major starter, considered a liability for a high major
5.4 C Typical "2-star" depth player. Sees some special team action. Would be a bad starter on a mid-major.
5.3 C- Liability player with obvious flaws in his game
5.2 D+ Preferred walk-on
5.1 D Walk-on
5.0 D- Guy in Section 31 eating fat-free pretzels
4.9 N/A Not ranked (shouldn't count in ranking)

The way to calculate this after, by the way, is to calculate the difference between the RR and the McBean numbers, then total that (keeping the departed players out of the equation).

So for 2004:

Name Pos Rating Grade McBean Grade Diff
Max Martin RB 5.8 B+ N/A N/A N/A
Roger Allison RB 5.6 B- N/A N/A N/A
Keston Cheathem WR 5.6 B- N/A N/A N/A
Marques Walton DT 5.2 C- N/A N/A N/A
Adrian Arrington WR 5.8 B+ A+ 6.1 0.3
Alan Branch OL 5.8 B+ A+ 6.1 0.3
Mike Hart RB 5.7 B A 6 0.3
Jamar Adams DB 5.7 B B+ 5.8 0.1
Chad Henne QB 6.1 A+ A+ 6.1 0.0
Morgan Trent WR 5.8 B+ B+ 5.8 0.0
Charles Stewart DB 5.7 B B 5.7 0.0
Jeremy Ciulla OL 5.6 B- B- 5.6 0.0
John Thompson LB 5.6 B- B- 5.6 0.0
Grant DeBenedictis OL 5.5 C+ C+ 5.5 0.0
Tim Jamison DE 6.0 A A- 5.9 -0.1
Chris Graham LB 5.8 B+ B 5.7 -0.1
Will Johnson DT 5.9 A- B+ 5.8 -0.1
Alex Mitchell OL 5.9 A- B+ 5.8 -0.1
Michael Massey DE 5.9 A- B 5.7 -0.2
Chris Rogers LB 5.8 B+ B- 5.6 -0.2
Doug Dutch WR 5.9 A- B- 5.6 -0.3
Brett Gallimore OL 5.9 A- C+ 5.5 -0.4
5.75 B+   5.77 -0.5

First, on the N/As. I chose to leave the N/A McBeans in the Rivals class because I wanted to use the Rivals class in it's final form. That's what generated the team ranking and that was the starting point. If we hack out the players who later transfer, I would have to theoretically circle back to every team and adjust their team ranking to get a new starting point - not possible. The debate still rages if my treatment is to rate a Dann O'Neill a ** star and punish the staff for recruiting poorly, but then that exacerbates the "two-star problem." I just remove them from the denominator of the McBean average, which serves to nuetralize those players. Still open for debate.

As for you beautiful chart above, I weep at it's niceness, but have to voice the same objections as above. In addition, we very much have to balance the McBean numbers with the Rivals numbers, which is discussed in the definitions post and would result in over half the players above being downgraded. Give that a thorough read (through the comment section) and see what you think.

First, I want to again assert the utter awesomeness of this whole project, and whatever you decide with I will be totally on board!

I think that even though class losses are part of a class's ranking, attrition is part of college football. Some guys will leave because they didn't work out. Some guys will bolt during a coaching change, or because their diaper wasn't getting changed, or what-have-you.

That's not the question you're asking, though.

The purpose of your rating system, if I'm interpreting this correctly, is to understand if RR is getting MORE out of his recruited classes than Carr, correct? The hypothesis, summed up as the "Pat White Factor," is that even if RR classes do not rate as highly on Rivals, they will end up just as, if not more productive than Carr classes.

This needs to be clarified to mean one of two things. Either:

a.) Those players who stay in RR's system will end up more productive than those who stayed in Carr's system
b.) Those recruits that RR brings in are better scouted than those that Carr found, and will thus become better all-around players.

If it's a.), we're testing the Barwis and RR scheme effect and whatnot, and in that case you should throw out all N/A's. If it's b.), then we should mark their career progression post-Michigan if there is any to mark, and those like Antonio Bass get "INC" and are removed.

Thanks for the kind words.

Since I want both a) and b), I think we need to review every N/A. That will be the next post.

Whoa. I should be angry with your for rolling in this late in the game with such a radical proposal that has merit. LOL.

We did discuss and rule out the RR because we were having trouble agreeing on ratings from a 5 point perspective. There was so much subjectivity that trying to slice it finer seemed to have diminishing returns. Also, two key data sets would be unavailable (2002 and 2003), giving us only one finished class.

What do you think in light of these objections. I very much like your proposal, though.

eh, we do what we can.

Personally, I'd go back and make up RRs for 2002 and 2003. Shouldn't be too hard, right?

Name Pos Stars RR
Gabriel Watson DT 5 6.1
Jason Avant WR 4 6.0
Steve Breaston ATH 4 5.9
Matt Gutierrez QB 4 5.9
Larry Harrison DT 4 5.9
Darnell Hood RB 4 5.9
Mike Kolodziej OL 4 5.8
Quinton McCoy ATH 4 5.8
Pierre Rembert RB 4 5.8
Carl Tabb WR 4 5.8
Jeremy Van Alstyne LB 4 5.8
Willis Barringer DB 3 5.7
Rondell Biggs DE 3 5.7
Mark Bihl OL 3 5.7
David Harris LB 3 5.7
Rueben Riley OL 3 5.7
Brian Thompson LB 3 5.7
Tom Berishaj OL 3 5.6
Kevin Murphy TE 3 5.5
Obi Oluigbo LB 3 5.5
Paul Sarantos LB 2 5.4
5.76

Damn, you are like a data machine.

I will throw your model out to the public for comment? Public? Please comment.

Name Pos Stars RR
Prescott Burgess DB 5 6.1
LaMarr Woodley LB 5 6.1
Shawn Crable DE 4 6.0
Leon Hall DB 4 6.0
Ryan Mundy DB 4 6.0
Jerome Jackson RB 4 5.9
Adam Kraus OL 4 5.9
Jake Long OL 4 5.9
Quinton McCoy DB 4 5.8
Will Paul TE 4 5.8
Jim Presley LB 4 5.8
Clayton Richard QB 4 5.8
Jeff Zuttah OL 4 5.8
Anton Campbell RB 3 5.7
Garrett Rivas K 3 5.6
Pat Sharrow OL 3 5.5
Brandent Englemon ATH 2 5.4

David Harris earned that 5th star at Michigan, and his NFL career continues to bear that out imhe.

Also, im tempted to think Jamar Adams deserves a 4th

We had quite the debate about Harris, but we exclude NFL career from our metric only considering the NFL draft as that was the professionals appraisal of how well you did in player development. It's close, though, as he went in the high second round and was a stud at Michigan. Others point to his slow development as a reason for keeping him at 4.

Adams went undrafted. Was he a multi-year starter? One vote for Adams as a 4 star.

even if you tone down the Henne/Hart love. Branch, Johnson, Arrington, Adams, Trent and Jamison alone are a very nice core group. you need about 5 starters per class and all the names mentioned were probably above average BCS starters. from '02 to '08, no M recruiting class got more total starts, despite both Arrington and Branch leaving a year early.

should Dutch and Adams be 3 stars.

Dutch never did anything - huge recruiting miss, forced to switch positions as an upperclassman because he was buried on the wr depth chart.

Adams was a solid if unspectacular starter, he contributed to the team.

Edit - to be more specific - ding a star from Dutch and then give it to Adams.

ding a star from Max Martin - but looking back on his stats he played a lot more than I thought before he transferred. Even had a 100 yard game.

Remember our metric, JHS. A functional backup at Michigan gets a third star. You may argue he wasn't functional, but he stuck around and I recall seeing him in the two deep once or twice in his senior year. We will survey on him.

Adams went undrafted, but he is, I admit, a borderline case. Does anyone remember how many years he started?

he was a second team AA by the AP and other less relevant outlets. thats top 2-3 MLB in the country.
and jamar was 2x All Big 10 second team

EDIT: meant to put this down under my previous comment

Great post! Up arrows for all!