Thanks for the in-depth insight and great analysis. That must have taken quite a while to compile.
here's one vote for "John Beilein's head in a Futurama jar"
Where are all these receivers coming from?
Many have noticed, and begun commenting, on Rich Rodriguez’s apparent stockpiling of receivers. This has been addressed a bit by Brian on the recruiting board, but I thought I’d go into a little more depth to explain why every RR is building up bodies at the position.
The depth chart looks awful large. Once the 2010 commitments step on campus, the receiver corps will include Martavious Odoms, Terrance Robinson, Jeremy Gallon, Darryl Stonum, Junior Hemingway, JeRon Stokes, Roy Rountree, Thomas Gordon, Ricardo Miller, Toney Clemons, Jeremy Jackson, James Rogers, Jerald Robinson, and likely Justin Feagin. That’s 14 receivers, with only Clemons in his final year. And don’t forget there’s still four scholarship players, Kevin Koger, Brandon Moore, Martell Webb and Steve Watson, at tight end. That’s 18 pairs of hands to feed. How is he going to do it?
Rodriguez’s offense, is, of course, different than that previously used at Michigan. One important distinction is the position of Slot Receiver. RR likes to have a scat-back type of player here with a slightly different skill set than your prototypical wideout. Size and leaping ability are secondary for a slot receiver to speed and agility. Circus catches aren’t as necessary, since most routes are short. The slot is meant to catch the ball in space, then make defenders miss. Odoms, Terrance Robinson, Jeremy Gallon, and, if he moves, Feagin fit this mold. Last year, Odoms pretty much dominated this spot, but that’s not typical of RR’s West Virginia offenses. Rather, Odoms got so many balls because once Robinson went down, he was the only man standing. Starting next year, I think you’re going to see a two-man rotation in the slot, with another man always ready to go for depth at an injury-magnet position.
In case this recruiting season didn’t clue you in, Rich Rodriguez is phasing out the Tight End position. I imagine, so long as he has Koger, the position will remain in the offense quite regularly through 2011. But come 2012, I wouldn’t be surprised to see just one or two Tight Ends on the roster for a change of pace, or goal-line situations.
That leaves Stonum, Hemingway, Stokes, Rountree, Gordon, Miller, Clemons, Jackson, Rogers, and J. Robinson. For one, we can imagine Gordon moving to linebacker, although this isn’t a given; since he was offered as a receiver, and unlike Feagin there hasn’t been mention from the coaches as to a positional change, I’m gonna consider him part of the receiver corps.
That gives us 10 pure receivers on the 2010 roster, all of them options. First of all, you’ll see some redshirts on some 2010 freshmen, probably J. Robinson and Jackson. Now we have eight. Eight is still a lot. And eight is actually what we’ll need!
THE SPREAD’S NOT JUST FOR RUNNING
Rodriguez’s offense at W.Va. was mostly about the run, particularly once the astounding legs of White and Slaton and Devine arrived. But I’m going to postulate, based on the recruiting focus since he got here, that plans are to make the passing game a greater part of the expected offensive output.
The formation I believe we’ll see more and more from Rodriguez will be the one-back, 4-wide. This includes a slot, and three wideouts. In a running-based spread, the wideouts head out on routes designed to open up space for the tailback, quarterback or slot receiver to function in. So long as they are all threats 1-on-1, they have to be guarded.
But other programs that use the spread have done a better job incorporating receivers. Brian commented this year that it sometimes seems like the receivers were running random routes – anything to get the secondary away from the ball. This was highly ineffective, especially since we didn’t have a quarterback who could get the ball to these guys.
I trust Brian in most things, but from everything I’ve seen of RR, if his receiving scheme was basically saying “get open,” then it wasn’t by design so much as he had other areas to focus on.
TALENT ATOP OF TALENT
This year, you can’t fault RR if the passing game wasn’t his primary focus. New team, new scheme, yada yada, but the talent really wasn’t there. Threet didn’t have enough time to pick apart a defense, nor was the redshirt freshman prepared by experience to fully utilize their talents. There wasn’t a true go-to receiver as we’ve had in past years. Greg Mathews was the most trustworthy pair of hands. After that, Stonum was a true freshman and played like one. Hemingway had mono. Roundtree was waiting for his muscles to grow onto his 6’3 frame. Clemons had catching issues.
There’s a progression formula I’ve been using to determine a player’s expected growth in value to the team:
Adjusted Star Rating * [1/2(years-in-school / years-in-school-plus-1)] plus (Adjusted Star Rating / 1.6)
(The “Adjusted” part means I change their star rating once we see them on the field. For recruits, I just use their star rating.)
Bigger jumps occur earlier in a player’s career. What we end up with is a level of expected performance based on their talent and their year, which is roughly equivalent to the familiar star rating system. A 2.00 player is what you’d expect for a typical starter at Indiana or Northwestern. A 3.00 player is a typical starter at Purdue or Michigan State. A 4.00 player is what you’d expect from a 4-star recruit in his 4th season. Over 4.50 is an All-Big Ten performer. Over 5.00 is a 1st round draft pick.
Here’s the receivers RR had available to him in 2008:
G. Mathews – 3.83
J. Hemingway – 3.50 (out for season)
D. Stonum – 2.81
T. Clemons – 2.63
Z. Babb – 2.63
R. Roundtree – 2.50 (redshirted)
L. Savoy – 2.50
J. Rogers – 2.19
Even though there’s talent there, it’s young talent. For Big Ten, that’s average. For Michigan, it’s mediocre.
Here’s projected 2009:
G. Mathews – 4.00
D. Stonum – 3.94
J. Hemingway – 3.83
R. Roundtree – 3.50
T. Clemons – 2.88
J. Stokes – 2.81
L Savoy – 2.56
C. Gordon – 2.50
J. Rogers – 2.40
That’s a huge difference. Your top three guys are expected to perform at or near what you’d expect from a senior 4-star recruit, whereas last year we had one guy near that level, and the next was below average.
From here, it’s all uphill.
D. Stonum – 4.31
J. Hemingway – 4.00
J. Stokes – 4.00
R. Roundtree – 3.83
C. Gordon – 3.50
R. Miller – 3.13
T. Clemons – 3.00
J. Jackson – 2.81
J. Rogers – 2.50
J. Robinson – 2.19
D. Stonum – 4.50
J. Hemingway – 4.10
R. Miller – 4.38
J. Stokes – 4.31
R. Roundtree – 4.00
J. Jackson – 3.94
C. Gordon – 3.83
J. Robinson – 3.06
And if you think that’s the only part of the passing game that will improve, look at what happens to our quarterback rating in that time:
2008: Threet/Sheridan – 2.27
2009: Threet/(Forcier/Robinson) – 2.69
2010: Forcier or Robinson – 3.50
2011: Forcier or Robinson – 3.83
The offensive line, too, will see a marked progression from about 3.20 to 4.20.
So in the years to come, Michigan is going to be stocked at receiver. In the 2011-2012 seasons, it is very conceivable, barring major transfers and losses, that RR will have at least four and as many as seven superb options at wideout. There will be a considerably better quarterback, protected by a considerably better offensive line. Is all this talent really just for show, or is there something more?
THE POSITION OF WIDE RECEIVER IN MICHIGAN’S OFFENSE, 2011 TO 2012
I have to imagine that Rich Rod knows what he has at these positions, and that Devin Gardner (who, if his 5-star is for real, would surpass Forcier or Robinson by 2011), is being made aware of it. The question remains, however, how do you utilize all of it.
The answer is a Spread and Shoot.
Look at the roster for Texas Tech’s Air Raid offense:
Missouri, too, had 16 receivers. Texas carried a ton. Florida and Northwestern had ‘em coming out of their ears. And note that Michigan actually carried 19 on the roster last year, though nine of those were non-scholly walk-ons.
Oklahoma, who uses a more Pro-Style offense, had considerably fewer.
Now, the spread has at most five receivers on the field. But in order to keep using the entire down-field as a threat, it's never the same guys. They rotate...a lot! For Michigan this year, however, the rotation wasn't there. If you sent Mathews and Stonum running in circles for five plays, you'd end up with Babb and Rogers. If I'm an opposing defense, and I've got to cover James Rogers down field, with Sheridan under center, then hell, I'm gluing a 3-star cornerback to him and telling the safeties it's backfield hunting season.
If RR has any specific plans for this kind of team, I imagine those plans are more vertical than anything college football has seen for awhile, and well more than anything we’ve imagined.
Most teams today – and we saw a lot under Lloyd – send a man on a deep route as a matter of course. But there’s a drawback – if you’re sending your Super Mario deep every play, he’s going to either let up on the gas, or wear himself out in three plays.
Go try sprinting 45 to 60 yards downfield, juking and turning various angles. Now hustle back to your starting point, and do it again. Repeat six times. Have someone whack you or knock you down a few times while you’re at it.
Two things will become apparent very quickly:
(1) Darryl Stonum is in much better shape than you, and
(2) huff....huff .... huff ..... there .... is. ....no...... huff ..... f’in .... way ...... someone ..... can ...... do ..... this ... huff .... 26 times in a row!
(the Barwis pit is over there, by the way. Help yourself)
I didn’t even ask you to out-jump someone and catch a pass.
BUT THERE'S NO "PASS" IN "RICH RODRIGUEZ"
What these spread offenses do a lot of is substitute. In a year with mediocre talent at receiver, horrible talent at quarterback, and a marked and stated preference for the running game, Rich Rod had 15 players catch a ball last year. Of course, many of these are running backs. So here’s another stat: James Rogers, the 8th man on the wide receiver depth chart, appeared in 9 games. Zion Babb was in 6. LaTerryal Savoy caught four passes all year, but lined up at receiver for 11 of the 12 games. Two safeties and three cornerbacks also lined up for the offense at times.
Those are slightly above what you’d expect from a Lloyd Carr team. But that’s a lot of receiver substitution for a team starting a walk-on QB and his noodly appendage.
They're also, by the way, WAY above what you'd expect from a Rich Rodriguez team, too. Clemson under RR utilized half as many receivers. Tulane spread the ball less as well. And West Virginia, as legend tells, was the runniest of the runny.
Yes, and you can't get Ricardo Millers to go to any of those schools, either.
In his previous gigs, Rich Rodriguez was very good at maximizing offensive output by maximizing certain positions. This is great at a school where you have to make recruiting decisions early and often. W.Va. can have a big season and be a player in the national dual-threat QB sweepstakes, but they were never going to be the kind of school that's on every kid's list in the country before they even get a call. RR didn't come here to turn Michigan into West Virginia. RR came here to further his career, to do better than he did at West Virginia. There's not much further you can go with the running game than he had with Devine and Pat White. What's wrong with imagining that he actually has designs on creating college football's ultimate offense?
The offense, I think, is going to make considerable more use of its receivers every year between now and 2012. The plan, as I see it, is to not just spread the field horizontally with a 4-wide alignment, but spread it vertically by having at least two receivers who can’t be left one-on-one going deep into the secondary on every play. It’s Terrell and Walker all over again, except while Stonum and Hemingway catch their breath, Ricardo Miller, JeRon Stokes and Roy Roundtree are doing the damage. Even if the ball doesn’t go that way, the simple necessity of covering multiple deep threats will keep the safeties back, and open up some space for the slot receivers, the crossing routes, and, of course, the slippery quarterbacks and smurfs-with-jetpacks running backs.
It makes too much sense not to. It’s the purest ideal of the spread, only realized here because Michigan can actually get enough talent so that the 4th or 5th receiver on the depth chart is worth double coverage. It’s the perfect marriage of the tried-and-true Lloyd Carr concept of maximizing talent differential, with the Rich Rodriguez ideal of making the defense cover the entire field, then beating them with speed and specialty-type players.
That’s why, at least in my opinion, Rich Rodriguez has stockpiled so much talent at the Wide Receiver position.
That, and because if you’re after a 2010 or 2011 5-star quarterback, already being stocked with targets for him makes an awfully good selling point.
But that still doesn't mean they're going to all be wracking up 1,000 yards. To that, I recommend reviewing the comments these guys made when they signed. They're not "I'm going there to be the next Braylon Edwards." If there's a theme to any of it, it's "this program is on the way up and I want to be a part of it!" He's getting guys who want to win championships; okay, every guy wants to win championships -- but he's getting guys who are picking their school based on where they think they can win championships.
And if you ask me, I believe RR thinks that way too.
Thanks for the in-depth insight and great analysis. That must have taken quite a while to compile.
Great diary entry, although your improvement method obviously assumes that none of those players are busts. I'm a little skeptical about Darryl Stonum, for one.
Even if some are busts, this list provides a starting point. Take out a third of these guys (or whatever) as being completely useless, and you still have a nasty receiving depth chart and a good year-by-year progression.
You are also going to have some of them boom!
The more lottery tickets you hold the better your chance is at winning. Same with any position on the football field; in college, as has been mentioned, it is more difficult to deal with elite players at skill positions. I think part of RR's recruiting is trading in FB's and TE's for extra chances at grabbing elite skill players. As has been mentioned you get the added bonus of forcing defenses to use "Johnny Sears" to cover Ricardo Miller (the up and coming) or your solid number 3 or 4 wr. We have seen who wins that battle.
Wow, an excellent read. Keep up the good work. USC has done this to an effect, in that they've stockpiled so much talent at WR, it's made it near impossible for opposing teams to succesfully defend them over and over and over. They continuously rotate 5 star talent in and out, receivers who are all 6'5" with 4.4 40 speed.
Many teams don't have two lockdown corners, let alone one, and most of teams have to shift more defenders into the backfield to cover all the deep threats, opening up holes for USC's running backs and tight ends to make plays in the open field. I think this is one reason why Carroll's system is so good. In the NFL, because of the draft rules and salary cap restrictions, teams can't simply stockpile WRs. They'll usually have one deep threat and a few other solid receivers. With USC, every receiver is a deep threat and a super, highly rated athlete; Carroll can recruit as many as he wants. It's just too much for college teams to handle.
This is exactly what happened in the Rose Bowl. Penn St played pretty good defense, but they couldn't come close to covering all of the USC wide receivers. No surprise really, not many teams have enough great DBs to do so.
Penn State made the mistake of playing a three deep zone against USC and Sanchez made them pay for it. If you're going up against a team with those types of weapons you need to play high risk/high reward defense and man up and blitz in order to achieve pressure. If you're going to sit back on your heels you're going to eat it.
Richard has a history of successful adaptation. It shouldn't be a surprise to people when he manages to redefine himself yet again in Ann Arbor.
Agreed. A truly great coach can adapt and best utilize talent. I think Rich has found a QB with great accuracy (T-Force) and envisions a passing attack in the coming years, similar to some of the B12 teams we saw this past season. It won't happen overnight, but when you go 3-9, 6-6 doesn't seem so bad the following year. Then you're a hero when you win 10 games in the 3rd year.
Any diary that includes a resonably defendable mathematical model is A-OK in my book. Kudos to you, sir.
What a read! you sir, are an awesome writer/analyzer/hypothesis maker.
Kudos, good sir.
I'm not familiar with the progression formula you've presented, but it seems rational and is very intuitive. Do you have a link to where this was developed and presented?
I put together what was to be a post talking about how it's going to take until 2011 or 2012 for Michigan to be back to the top of the Big Ten, but that the talent for this was already there.
I showed it to Brian, and sent it to friends in e-mail, but I couldn't figure out how to get it posted, and the Diary died.
It was my own progression formula. Aside from Stats 402 in like 2001, I don't really know statistics, so I kind of blundered my way to it the hard way.
What I wanted to do was find a way of translating what a certain star rating meant on the field. I had a list of Michigan's recruits since 2004 compiled during my previous diary on Position Rank versus Star Rating as a measure of a recruit's potential. So from memory of each player, I put down what level I thought they were playing at for that season. The ratings were based on the star ratings, but if that player had been a senior, i.e., if this level of performance was put out by a 4th year player, what star rating would he have? Then I sorted them by initial star rating, took out the outliers (who defied expectations or underachieved), and then came up with a formula that met that progression.
The 1.6 constant is the multiplier by which a five-star player improves between his freshman season and his senior year. The rest of the formula just makes sure the progression is curved.
How do you get the "adjusted star rating" once the person has been playing? I assume you could not project into the future until that person has played. What variables go into the calculation? For example, do Braylon, Hart and Steve Brown to show the adjustment.
After (or during) the season, once we've seen how a kid plays, I can rate him against the others on the team at his position.
It's a rolling thing. After several years of running backs, we knew that David Underwood was at a certain spot. His progression was exactly what you'd expect from a 3.5-star. So was Jerome Jackson's. Underwood (3.50) won the starting job, and freshman Mike Hart didn't play until after Underwood was injured. Hart then beat out all the other uninjured backs on the roster. Because we'd seen Jackson for years by that point, there was no reason to down-grade him. He was a 3.5-star player, playing at a 3.06 level, exactly what he should be at in his career. But Hart was already by him. Well, a true freshman playing at a level above 3.06 had be a 5-star (3.13). I adjusted his star rating to 5, and that rating held throughout his career.
Column I being Adjusted Star Rating
Column G being the year of matriculation
B1 being where I put the current year because A1 had "Year:" in it.
All I can say is that I am deeply impressed, and I hope you are right. This offense will be a thing of beauty if it comes to pass as you envision it. I have to admit, I've been wondering about all of the extra receivers, and if RR was going to build another WVU spread, this just wouldn't mesh. He'd be stockpiling running QBs and running backs, not WRs. While we have decent RB depth, it's not nearly as deep as WR and probably no deeper than a typical pro set team.
I like the 1.6 as a fudge factor, well done.
Your theory does explain the smaller/quicker LBs being recruited to cover underneath routes of teams with 4+ athletic WRs. It is going to take getting guys like Gholston though to stop the run for teams remaining that want to pound the ball. A few more big DT's to plug up the middle would be nice.
Some nice golden ratio action there! I like...
You sir, are my hero. Not only did you think of a valid point, but you supported it with math. I think I love you.
...but like, as a friend, right?
I recall reading of an unnamed Big10 coach analyzing our loss to Ohio State in 2006, saying Ohio State's 3rd receiver (Robiskie?) was better than our 3rd cornerback. (After Hall and Trent, I can't even remember the 3rd cornerback). Having the ability to put 4 or 5 high level receivers on the field at the same time, with a good quarterback, line and running back, is very powerful. How many opponents will have 4 or 5 top notch cornerbacks? This is simple arithmetic, but it could be very effective. (Only us true fans could be this excited after 2008).
Just for kicks, we should put that formula up for some "open source" modification. For example, I don't think the improvement a player makes from year 1 to 4 is remotely linear for a receiver. It's more logarithmic, in the ballpark of y=10ln(x)+1.5
Mine wasn't exactly scientific. But it was curvy:
I think I misread your formula as being linear. It obviously is not.
A very interesting read. I like the analysis for sure.
1) Though this has been addressed already, I think your rating system is a bit off in assuming that Hemingway and Stonum and Roundtree are all going to have breakout years. As of yet they've all been highly mediocre, with Hemingway showing very brief flashes (and Stonum catching that one deep ball). With the limited amount we saw Stonum and Roundtree (not to mention the fact that they didn't have a QB who could get them the ball) I'd be surprised if they're much improved, at least to the level you suggest. Though I'd be satisfied with just one having a breakout year. A good general system, though.
2) Protection. For a ball to be thrown deep accurately, a QB needs time to allow his receivers to get down the field. I'm not saying we'll be making it rain all day, per se, but with 4 receivers and a small running back, I question the amount and quality of protection our QB will get especially in the face of heavy blitzing. So I'd say don't expect too many deep balls. Rather, I expect RR to implement a Texas Tech/Texas/Missouri-style spread passing attack with lots of short and intermediate passing. Add an athletic QB to that, and you've got a deadly offense.
Your idea of a marriage of Texas and West Virginia's style of offense sounds very exciting. One thing that RR has done at each stop he has had is to innovate. I don't think that stops here. You're going to see a new brand of RR and Michigan football emerging in the next few years, and it should be very fun to watch. Now if we could only get the defense to work...
Very thoughtful, well-written diary.
I just want to add one point that was not explicit in your analysis, although you touched on it a little. With all these WRs, teams will be forced to drop safeties and LBs back, which, combined with a running QB, can open up the field for those defense-breaks-down-QB-scrambles-for-a-big-gain plays I remember from the Troy Smith days (so frustrating). In other words, we can establish the pass in order to run.
Plus, how many of the receivers can be Percy Harvin-types, that regularly line up in the backfield, and take the option or flare in open space?
You should never let math-based facts get in the way of a good argument. That said well done. You certainly won't find that type of intelligent posting on an MSU or OSU board.
keep up the good work man
This post echoes what I've always thought about how 1337 Rich Rod's offense is, which is that:
1) The focus is on creating one on one matchups all over the field;
2) Winning as many of these matchups as possible.
Obviously, the more talent you have the more effective you will be at #2. In terms of the "RichRod + Michigan talent" equation over the long run, I can easily see us being the best offense in yardage and scoring on a regular basis. Without question, there are things RichRod's been looking to do over the years that he simply hasn't had the talent to execute.
That Ninja Football guy is right: we ARE going to be a machine!
But I think the statistical analysis is too arbitrary in that it assumes lack of injuries, busts, continued improvement (not plateauing), other issues.
Now if none of the above happen, yikes.
But I'm not sure. Namely
- A great long ball requires great *timing* too. I'm not sure you're going to get that shuttling WR's in and out
- A great long ball attack requires a great QB. You're only going to get that maybe 2/3 years at best.
- M QB's are going to be spending time learning read option. There's only so much you can learn.
- M QB's are going to be getting hit and hurt.
Lastly, UF also stockpiles talent all over the field and
especially at WR. At best, they run a balanced attack.
I think you've hit on a lot of good points here. You're dead on with the receiver stock piling and rotation. I believe Rodriguez said at one point that he likes to travel with 9 receivers which is a lot, and it was clear that he likes to rotate a lot of guys. However, I think the "random" pass routes may be a little off. I haven't had time to rewatch the games, but from what I remember, the down field passing game was based around selling 3 and 4 vertical. We had some success with this, especially with the tight end in the seam. Several times there were guys open in the seam, but Tacopants ended up with the reception. There were a couple things we did off of this. One was the slot guy reading the safety. If the near safety got over top of the slot(mostly Odoms), he would break off the route to the inside. This also worked several times, half of which were complete, and half of which ended with ball in the dirt. We also liked to cross the #1 and #2 receiver sometimes and send the slot down the sideline and #1 down the seam. However, we also ran the curl/out combo off this look. This is a pretty common route combination where the outside receiver runs a curl, the inside receiver runs an out or arrow to hold the flat defender, and the tight end runs a check release (check for blitzers) hitch to try and hold the hook to curl defender inside. In Michigan's case, the outside receiver sold the inside seam before breaking off on the curl. This worked very well against Notre Dame, and was effective in other games as well. Greg Mathews was great at this route. You also saw the wheel route with #2 or with a RB out of the back field out of this look. Against Illinois, a primarily cover 2 team, you saw the smash route concept. I think we did this out of both a straight drop and full roll. This route creates an over under read on the flat player where you send #2 on the corner and #1 on the stop or smash route. We hit a couple of the corner routes against the Illini, but I don't remember much success with it against the other teams. We had a pretty standard three step package with fades, slants, and hitches out of pretty much every formation. I think the slants disappeared in the middle of the season, but Threet was pretty effective throwing those against Notre Dame and Purdue. I was wondering why we didn't see more of those, but maybe teams were taking them away. There was also the one step screen game which just never seemed in sync. Anyway, I think that the passing game is going to continue to be based around hitting those seams and forcing safeties to defend the deep routes. With the more mobile qb's, you will also see play action off the speed option which RR loved to run at WV. You may remember this play from the bowl game against Oklahoma where the WV receiver was running down the field by himself. As far as returning players, I think Stonum has the best chance of being a big play threat on the outside. He just needs to clean up his technique which is true for just about every freshman receiver. But anyway, it was a good post. I always enjoy reading stuff about the passing game.
the number of followers of this blog who are mathematicians, engineers or statisticians. The enthusiasm of some here for your progression formula resembles what I imagine reaction would be at a Star Trek convention when Leonard Nimoy takes the stage. Myself, being merely a dumb Political Science/History guy, I appreciated your post quite a bit, only stumbling at the progression formula, and savoring more the theory aspects, which I hope are true in future. Thanks.
History with a French minor here.
Unfortunately I've yet to find a way to incorporate the reforms of Henry II into a discussion on Michigan football, but don't worry, I'll find a way.