Another (Alternative) Look at Returning Production

Submitted by DrMantisToboggan on February 20th, 2018 at 1:43 PM

With Bill Connelly recently releasing his returning production rankings (link below), I wanted to take an admittedly more simplistic look at returning production on our schedule and in the Big Ten. I've started with the offenses and might do a defensive version if this is well-received.


I think the best barometer for what a team returns on offense is the raw percentage of yards produced from the previous season. For the purposes of this exercise, I have factored in Shea Patterson's 2017 statistics as if he were to be declared eligible in the fall, because by most accounts he will be. Numbers for every team in the conference and on our schedule in the chart below, sorted by Total Percent Returning.

Team 2018 Returning Pass Yds/2017 Pass Yds 2018 Returning Rush Yds/2017 Rush Yds 2018 Reurning Rec Yds/2017 Rec Yds 2018 Return Off. Production/2017 Off. Production Total Percent Off. Production Returning
Michigan 2931/2259 1805/2310 2078/2226 6814/6795 100%
Wisconsin 2689/2689 3025/3121 2011/2689 7725/8499 91%
Purdue 3270/3270 1944/1970 2158/3270 7372/8510 87%
MSU 2793/2798 1614/2182 2289/2798 6696/7778 86%
SMU 3789/3823 2396/2397 1502/38/23 7686/10043 77%
Maryland 1721/1940 1881/1940 840/1940 4442/5820 76%
WMU 1978/1978 1469/2697 1584/1978 5031/6653 76%
NW 2873/3040 746/2264 1917/3040 5536/8344 66%


2473/2473 305/1810 1689/2473 4467/6756 66%
OSU 626/3679 2607/3405 3659/3679 6892/10763 64%


3736/3772 939/2212 1404/3772 6079/9756 62%
ND 2326/2326 1598/3501 1099/2326 5023/8153 62%
Illinois 382/2098 1082/1267 1527/2098 2991/5463 55%
Indiana 1252/3188 1476/1561 1493/3188 4221/7937 53%
Minnesota 0/1513 1371/2178 1186/1513 2557/5213 49%
Nebraska 192/3330 1345/1290 2297/3330 38347950 48%
Rutgers 676/1387 454/1765 741/1387 1871/4539 41%


Some Notes:

  • Obviously this shows how important it is that Shea be declared eligible. Without him, we return only 672 passing yards, which drops us to 67% returning overall (there's a slight chance that would cause Wilton to come back, which would put us at 76%). Some may point out that Shea accumulated his passing yards in a different system with different teammates, so it might not be accurate to simply transfer 100% of his production to our team for this exercise, and that is a fair critique. That being said, I think the larger point with including Shea's numbers is that Michigan is bringing in a guy who has started 10 games in arguably the best conference in the country - that is a huge talent and experience boost to an offense that is now plenty experienced outside of the QB position.
  • Wisconsin brings back nearly everyone from the 41st ranked S&P offense. The only major loss is obviously Fumagalli. Fumagalli was Hornibrook's go to, so we will see if that affects their passing game. They do get Quintez Cephus back from injury, however. 
  • Purdue loses their top two receivers, but brings back everyone else. Jeff Brohm should see more improvement this year and Purdue probably gets up to 8-4 or 9-3. A couple big boys in the conference need to make sure not to overlook them this year...or do overlook them and get upset, that's fine with me.
  • Michigan State returns a lot as well, but they are dangerously thin at RB. Michigan also returns a higher percentage of receiving yards than MSU, which Connelly believes is most indicative of offensive improvement. 
  • Ohio State's numbers are deceiving. They lost JT Barrett who had a great statistical year, which puts a huge dent in their returning passing yards and a significant dent in returning rush yards. However, in limited action, Haskins actually put up better numbers than JT, and, to make that even scarier, nearly 100% of OSU's receiving yards return. They lost far more on defense than they did on offense, but OSU's offense is going to be a problem this year as long Haskins transitions smoothly to full-time starter.
  • The poor get poorer. Minnesota, Nebraska, and Rutgers are the only teams on this list that return fewer than 50% of offensive production. PJ Fleck has no Quarterback returning that has thrown a pass, and Scott Frost barely has one at Nebraska. Meanwhile, I had no idea that last year's pretty bad Rutgers was actually very experienced. Rutgers is hoping that losing bad seniors is actually addition by subtraction (except they lost an impressive RB in Gus Edwards and the electric Janarion Grant, so probably not).
  • Finally, Notre Dame returns both QBs (one was okay, one was pretty subpar as a passer) who played last year, but returns less than half of both rushing and receiving yards, and are even thinner than MSU is at RB. Plus, this chart has not taken into account Offensive Line losses, which Notre Dame has suffered in losing two probable first rounders in LG Nelson and LT McGlinchey. I think they are going to struggle mightily to move the ball against us in Week 1. Penn State is in a similar situation with their QB (who is actually good) returning, but the vast majority of rushing and receiving yards departing, and they should see an offensive set back year as well.


Bill Connelly's Article:…



February 20th, 2018 at 2:32 PM ^

I think we can say with confidence that Michigan wins all but the PSU game last season if Wilton Speight (2016 edition) takes every meaningful snap at QB.  I don't see why Michigan can't beat most of the teams on its schedule this coming season if it gets C+ or maybe B- QB play.  And Patterson can provide that if he's eligible. 

As for Haskins at OSU, it will be interesting to see how he handles being the full time QB.  A lot of pressure can bulid up pretty quickly in Columbus if you have a couple of bad games.


February 20th, 2018 at 3:37 PM ^

Yes, and I think you can reasonably expect B+ from Shea, at a minimum.

Yeah I doubt that Haskins can keep up a 9+ ypa and 4:1 TD:INT ratio - if he can he’s going to win a National Championship and Heisman. From what I’ve heard in practice he’s been kind of a high risk high reward guy, so I expect his turnovers to go up a bit and I also don’t think he’ll be as effective of a runner as Barrett was.


February 20th, 2018 at 6:59 PM ^

Yeah, I think his passing stats will come back to earth a little when he plays an entire season, but I still think he will be very good. I think he reflects what Urban wants to do (or at least what I have surmised that Urban wants to do) with the offense, which is gear it a little more towards a versatile passing attack and make it a little less run heavy (at least less QB runs).


February 20th, 2018 at 5:13 PM ^

portray the benefit of bringing in Shea would be to also use his yards as the denominator.  If you're going to include them (and I think that's fair), the base yards shouldn't be our tire fire of last year's QB production.

Also normalizes the calculation to other teams that are bringing back their starters and are getting a ratio of 1 returning not a ratio greater than 1 (which makes no sense).


February 20th, 2018 at 6:06 PM ^

To your first point - you're right that would probably be better to reflect "returning production". I think I left it how it is now to fully illustrate the importance of adding a guy that alone, in only 7 games, threw for more yards than we did as a team. I think I will change it to make the denominator Shea's 2017 stats. Not sure there is a great way to measure the experience impact Shea has on next year's QB situation.


To your second point - it can actually make sense, though rarely, to have a returning field be greater than 100%. Mostly that would be in the case of an impactful transfer. In this data in happened in one other field - Nebraska rushing. Nebraska returns all meaningful RB yards, but also loses a QB that had a lot of negative yards, and it makes sense that, under Frost, their QB would not have negative rushing yards. 


February 21st, 2018 at 11:27 AM ^

than a returning All-American?  If we returned a 4,000 yard all-american QB, the returning production at the QB position would still only be 100%.  The weight on the overall formula would be greater than returning a 3000 yard QB because the QB production is a greater percentage of the offenses overall yardage, but you'd still end up with less returning production as a percent than you're getting here. 

And that's why it makes no sense to allow for a position to return greater than 100% production when evaluating returning production.

By using his previous years production for numerator and denominator, you're still measuring the effect of his experience because you're saying that 100% of the QB production is coming back (which is what you're doing for all other returners with experience, you're not projecting them to increase their production) and by using his production, you're making it a greater proportion of the overall offense.

The way you have it makes it an advantage to have a transfer coming in over a returning starter.  One could easily argue that a transfer's experience is less valuable because there is a transition and a learing curve, etc.



February 20th, 2018 at 6:23 PM ^

They'll be good. I think they're a 9-3 team this year. S&P had them at a bit lucky last year, saying they probably should've won 9 games instead of 10 (aka probably only 8 in the regular season because there wasn't much luck involved in that bowl game of their's). The thing that makes me more comfortable about that game is their schedule, their depth at RB and general inability to run the ball, and their offense. They finished with a worse offense than us in S&P last year, they have nobody behind LJ Scott, who has had various on-field and off-field issues, and before our game they play 5 bowl teams from last year (Utah St, Arizona St, CMU, Northwestern, Penn State), have 3 tricky road games (@Indiana, @Arizona State, @Penn State), and have a very difficult game the week before ours (@Penn State). I expect MSU to have at least 1 loss going into our game, possibly 2.


I'd put our victory at MSU close to 90%, if not above. Jim knows we need that one and is pissed that he's only won 1 of those 3 dumb games. That's a game we will have circled and take care of. 


February 22nd, 2018 at 1:09 PM ^

Arizona states coach is herm edwards,CMU and Utah state are both much less talented than them and Indiana isn't a tricky road game for seemingly anybody but us.  The only games on that list I see finishing within two scores are @penn state and Northwestern but they could certainly have lost by the time we play.


February 20th, 2018 at 6:43 PM ^

Their offense was dead middle in the B1G last year at 7th (92nd Overall). I don't think they will improve past us, although I think ours will improve a lot. I think we will get up into the 30s to have the 3rd best offense in the league, while I would guess that MSU's offense stays out of the top 50. They couldn't run the ball at all last year and they lose 2 of their 3 best backs and their best OL. Lewerke is a decent QB, and they have a few decent receivers, but I can't see their offense improving dramatically. They will continue to be a "bet you can't score 21" team.


February 20th, 2018 at 8:46 PM ^

To me, I don't think who the RB is matters much at MSU. It's kinda plug-and-play a la Wisconsin and Stanford. 

I think returning 4 on the OL will be big for them, Lewerke now has a full season of experience, and they return their top WR's and TE's. 

If Lewerke can be as efficient as Hornibrook, but more mobile, I think that'll make MSU very dangerous. It just depends on whether or not he turns the corner and becomes the next Kirk Cousins or will he always be a loose cannon that makes bad reads. 

I think their offense has as much potential as our offense. I would take our groups of RB/WR/TE every day and twice on Saturday over theirs, but they have a proven system in place that works in the B1G, and their players buy into that system (kinda like Wisconsin). 

And just for arguments sake, they gave up over 21 five times last year and went 2-3 in those 5 games:

W vs Minnesota and PSU
L vs ND, NW, and OSU

On the flip side, they scored under 21 six times last year and went 4-2 in those games:

W vs Iowa, Michigan, IU, and Maryland
L vs ND and OSU


February 20th, 2018 at 9:36 PM ^

I don't think that's true about their RBs though. They were 66th in Rushing S&P this year, 55th in 2016, 98th in 2015, 17th in 2014, 77th in 2013, 69th in 2012, 87th in 2011, and so on. They've been objectively good at running the ball in exactly one year under Dantonio - their Cotton Bowl year. They weren't even good running the ball when they had Leveon Bell. 


I would take every position group on our offense over theirs. We have better athletes at receiver, more productive RBs, better TEs, if Shea is eligible we have a better QB. Our OL is clearly better at run blocking, but theirs had an edge in pass pro last year. 


Ultimately I think they are a sound team, but I think the difficult schedule gets them. Theirs is a tougher schedule than ours. Eight of their first ten games are against bowl teams from last year, with a brutal stretch of Northwestern, @ Penn State, Michigan, Purdue, at Maryland, Ohio State. There's no way they come out of that stretch with less than 2 losses. 


February 21st, 2018 at 8:35 AM ^

And also thanks for being one of the consistenly positve voices here on the board.  Not in a "Pollyanna" way but rather when presented with data you always see the good first and the negative second.  Lord knows I wish I could (as you know I'm a bit pessimistic right now) but I truly appreciate voices like yours calling attention to the reasons to be optimistic (and there definitely are logical reasons to be so) cause it's sorely needed right now.

Don (yeah that's my real name)


February 20th, 2018 at 7:22 PM ^

if he is eligible and stays healthy the entire season it will be a dramatic impact on the season. The B1G East is going to be brutal next year and our schedule is not going to help. Herbert needs to make a big difference in how our OL can make big strides of improvement this offseason. An interesting column to add would be the number of returning starting lineman for each program. Thanks for the data and perspective.


February 21st, 2018 at 9:24 AM ^

It is easy to get discouraged after the losses at the end of last season, but prospects for next season are not as dark as some insist. While it is true that we put up bigger numbers on offense against weaker teams, is that not the case for most, if not all, teams? Their stats could be seen as similiarly inflated.

I'll be interested to see your data on defense, but prospects there look good: not only do we lose few players, but we gain depth from the young talent that has gained a year's experience. That should help reduce the fourth-quarter fatigue that allowed other teams to score on us. And of course a more productive and consistent offense will give them more of a breather. Let us hope Shea's transfer is approved, or that the younger quarterbacks can step up. (Both would be nice.)


February 21st, 2018 at 12:21 PM ^

Interesting data, I wonder how strong of a correlation there'd be if you used this to predict next years offense. I'm guessing there'd be a decently good trend to something like:

2018 Total Yards = 1.25 * (Returning 2017 Yards) + 0.25 * (Lost 2017 Yards)

1.25 and 0.25 are complete guesses I'm throwing out there, but would essentially be saying returning players tend to increase production by 25%, while 25% of yards lost to graduation/early draft entries are replaced by freshman.