Turnover Turnaround Take Three

Submitted by Brian on May 17th, 2011 at 12:04 PM

The most persistently wrong thing I was wrong about amongst the many persistently wrong things I asserted about Rich Rodriguez and his bite-sized Era at Michigan was:  "that turnover margin is going to be a lot closer to zero this year." Or words to that effect.

I should have been right, or at least in the general ballpark of right. Rodriguez's West Virginia teams were consistently in the black. Turnover margin is so weakly correlated from year to year that Phil Steele annually puts out a "turnovers = turnaround" post highlighting teams with double-digit swings in either direction so that he can predict against the teams with big numbers and in favor of the teams with little tiny ones. Here's your favorite team:

Going Up TO

And here's how that worked out:


Amongst the many things that got Rodriguez fired, the persistently huge negative turnover margin is neck and neck with transferpalooza and program alum white blood cells for second place.

So these days, running across stats like these

running the correlation between one year's turnover margin and the next, I found that the correlation was a mere 12%. That's still something, but it's clear that for most teams, the turnover margin they enjoy one year has virtually zero predictive value for the turnover margin they will enjoy the next year. That means that on average, teams with substantially positive margins will see major decline in margin the next year, and teams with substantially negative margins will see major improvement the next year. A team with a -10 turnover margin in 2009, for example, would have an expected turnover margin of -1.2 in 2010, an improvement of nearly a full turnover per game!

…make me want to bang my head against the wall. Of course Michigan would be as large of a chunk of that correlation as possible, and of course they would be on the negative side of things. Thump. Thump. (The only thing worse than defying this correlation is defying the correlation between turnovers and wins: GTP points out Georgia went from –16 to +10 and still finished two games worse than they did last year. That would seriously harsh my buzz if I was a Georgia fan.)

I have tried to make the world make sense and this is what I've come up with:

  • The correlation is so low because quarterbacks are so important. The general path of a quarterback is: bench, horrible interception-laden youth, cool as a elderly cucumber, repeat. When you have a senior quarterback you are likely to have a good turnover margin and guaranteed to not have that guy back next year.
  • The same goes for everywhere else, albeit to a lesser extent. Good defenses tend to have seniors on them and tend to not have many returning starters.
  • Michigan eschewed the general path in favor of freshmen or equivalents. Rodriguez never started the same guy two years in a row, and while that turned out to be the right decision it meant three straight years of horrible interception-laden etc.
  • When your defense is so, so bad you escape TO margin gravity. AKA GERG. Also, what's the one thing that consistently generates turnovers? QB pressure. Michigan finished 98th in sacks and loved them some three-man rushes.

So while year-to-year correlation is low across college football, if you correct for experience—especially at QB—and maybe lack of prominent walk-ons/converted WRs in the secondary that 12% would be significantly larger. Michigan's program got so messed up that they stopped participating in the circle of turnovers*. Instead they laid at the bottom of the national rankings, a corpse dragged down by redzone interceptions.

Seriously, This Time It's Serious

Um… so… there's the above theory. And then there's Denard Robinson, Michigan's first returning starter at QB since Chad Henne. And then there's Greg Mattison and a defense that uses Craig Roh as a pass rusher instead of a clunky linebacker. There are a bunch of returning starters everywhere, including four guys on the offensive line.

I'm going to be the guy who puts his hat on a stick and pokes it out of a trench to see if there are any snipers around, again: Michigan's turnover margin should scream towards zero this year. They've got gravity on their side and many things besides. Also, Brady Hoke's miraculous digestive tract will move all that Tyler Sash wackiness to Ann Arbor.

This is the year Michigan has a mediocre turnover margin. Believe.




May 17th, 2011 at 12:17 PM ^

Is this like that Blackjack strategy where you just keep doubling your bet until you win? You gotta come out ahead eventually, errrrr ... right?


May 17th, 2011 at 12:29 PM ^

we have a returning starter at QB, for his second season, not 3rd or 4th, which I think is an important distinction to make 

He is returning to an offense he doesn't know, will be primarily under center in taking snaps and very likely trying to pass in a scheme that does not leverage him as a running threat to open up receivers, rather it would seem he is going to need to make more precise throws and have much better timing than he showed last year.

We have a more experienced D-Line, however we are still depleted in the DB department in both talent and experience, so the QB pressure may result in more errant throws, but I'm not sure we are going to be a pick-happy defense just yet. 

Add to that a gaping ?? at the RB spot where a good portion of the returning players aren't necessarily known for having a kung-fu grip and they are going to perhaps be pushed by some younger RBs that are, well....young. 

Toss up, we shall see.




May 17th, 2011 at 2:01 PM ^

The change in scheme & coaching may negate Denard's added experience.  And if he's hurt (again) the backup is another inexperienced underclassmen (again).

The defense is returning many starters but a) is still young in the secondary, b)is very inexperienced at LB and c) also has  new schemes and coaching.

I'd guess the turnover margin will improve, but may still be negative and that much of the change will come down to simply fewer total plays (due to Hoke/Borges doing some old-school we-need-to-control-the-clock type play calling and fewer big plays in general on both sides).

Three things, other than regression to the mean, that may help matters out:

1.  More conservative play-calling with the POWER mentality.  Less reward, but less risk (i.e. turnovers) from the offense.

2.  A very veteran OL (4 guys who are 4th/5th year plus Lewan) that also happens to be pretty good.  That's going to keep defenses at bay and give Denard more time than he's ever had in passing situations.

3.  The home games to start the season mean added familiarity, comfort, clairty of mind, and less crowd noise.  That's a nice way to transition into a new scheme.

Finally, Tate Forcier was a risk taker - flat out.  It didn't always mean turnovers but the guy wasn't afraid to take chances, for good and bad.  Denard and Devin (from what little we know of him) both seem to be more conservative.  Addition by subtraction in terms of turnover margin.



May 17th, 2011 at 2:38 PM ^

If we look at our probable starting DB's it looks like the following:

CB1: Troy Woolfork/Woolfolk/ T-wolf - RS Senior

CB2: JT Floyd/ Avery - RS Junior/ So

SS: Kovacs/Grit - RS Junior

FS: Thomas Gordon/Marvin Robinson - RS So/ So

I think our depth is more the issue this year. If we stay healthy, and our 4-man line can somehow out pressure last years 3 man line, I think our coverage should be much better.

Of course there's always "Never Forget"


May 17th, 2011 at 3:48 PM ^

Besides Woolfolk and Kovacs, the projected starters (Avery & Johnson) are class of 2010 guys .  Two upperclassmen in the starting secondary besides what should be RS Freshman, I'd argue, is fairly young, though not as bad as last year, obvs.

Furthermore the depth, other than Floyd, is underclassmen as well - Robinson, Furman, Talbot plus whatever freshman play - likely at CB.  T.Gordon will play (reportedly as the nickel back) but is a position change and still technically an underclassmen as well.

Agree we'll be better, but you'd expect upperclassmen to make up 3 of the 4 starters plus some decent depth behind them.  You're talking about Woolfolk, a walk-on, and JT Floyd - not exactly a recipe for a lot of INTs.


May 17th, 2011 at 4:22 PM ^

I think that we agree that it is backup youth that is our main issue.  Woolfolk is a very seasoned senior.  Avery started 5 games last year with Floyd (an upper classman) backing him up.  Kovacs (lets drop the walk-on tag, he's started and played well for two years) will start as an very seasoned upper classman. Which leaves us with a first year starting Free Safety, which, like, yeah, see Michigan Free Safeties of last decade.

If we stay healthy here, we have a good chance to be past mediocre to good (not great, but good is a huge step). I also think that defensive line pressure is a huge factor in secondary turnovers.  If our line is applying consistant pressure with 4 rushing and occasional blitzes, an uncomfortable QB is going to toss up a lot more bad tosses. I think we have a good chance to move back to the median on defensive turnovers.  


May 17th, 2011 at 6:21 PM ^

My statement up front was that despite the experience, they're still young.  Avery and Johnson have still arrived on campus last August.  When was the last defense (prior to Rodriguez) that started two 2nd year players in the secondary?   How often does that happen? How often do PSU, OSU, and Wisconsin do that?

You could even argue that with Floyd, Simmons (a walk-on like Kovacs), and Gordon as our top backups the depth is no younger than the starters.

My point isn't to say the defense will be bad.  I think we'll be way better.  But - we're still pretty young in the secondary.  You're talking about 2 sophomores, a walk-on, and a 3-star legacy recruit coming off a broken leg.  Lets keep expectations regarding a ball-hawking secondary in check. 

Now, if a year from now you're telling me about C.Johnson, M.Robinson, Avery & Countess returning as experienced players with veteran backups in Floyd, Kovacs and Gordon I might start getting excited about the secondary driving an improvement in turnover margin...


May 17th, 2011 at 2:29 PM ^


I was always taught your footwork as a QB is critical in limiting interceptions; perhaps more so than decision making ability.  Overthrow/underthrow, velocity, time to allow the play/receivers develop, and not leading the D (via body position) into the route are all dependent upon footwork.  Unfortunately, this is where Denard needs to develop the most in his transition.  He is a hard worker and I have faith in him.

Mattison's 4-3 under D is aggressive and hopefully opportunistic towards causing more turnovers.  



May 17th, 2011 at 12:29 PM ^

Interceptions are what worry me.

For all the dilithium he has, this is Denard's chink in the armor.

He makes some real bad decisions from time to time. I hope Al(l) will be fine.


May 17th, 2011 at 12:36 PM ^

I do think that Denard's decision making is his biggest flaw.  That flaw tends to be more exaggerated when he is facing pressure (duh!).  So, to some extent, Denard may always turn over the ball, and simply trying to bring down the number of turnovers to something remotely acceptable will be the goal.

HOWEVA, iif we ignore 2009, when Denard didn't even know the playbook and should have been redshirting but our team lacking any backup QB, last year was Denard's first season.  He was the equivalent of a redshirt freshman.  Remember Jimmy Clausen's freshman campaign?  He threw picks more often than TDs.  Over his career, those picks went way down.  So, naturally, I expect Denard's TOs to decrease. 

Also, my personal feeling (please no debate about RR, the spread, the state of our country's government, gun control or any other hot button issues) was that our 2010 offense placed far too much burden on Denard.  Whether this was by design or as a result of not really having an effective RB to share carries, we will never know, but the bottom line was that Denard either threw or ran on a huge % of plays.  With Borges' professed plan of cutting down the # of Denard carries and putting more of a burden on the RBs, I would expect Denard's TOs to decrease markedly.


May 17th, 2011 at 1:49 PM ^

I understand your point and I can see some danger signs, but I fight this point in every thread so I can't stop here.   What Denard was last year was not a finished product.   He was starting for the 1st time and had an incredible burden on his shoulders.   Turnovers also tend to happen when you are behind because you are forcing things.

Do you think that if you were in Denard's shoes you could have said aw hell I'll just throw this ball away the defense will stop them and get it back to me?  

Denard made bad decisions...so do a lot of qb's when they 1st start out.   Peyton Manning threw  a bunch of int's his rookie year as well.  Remember K. Cousins choke artist in 09?  How about Pick 6 Ricki Stanzi in 09?   Tolzien killed Wisc in 09 with crucial turnovers against OSU etc...

Let's give him some time before we define him.

Mike Martin breakdown in 08....not strong enough at the point of attack.

Mike Martin breakdown in 11......pain.





May 17th, 2011 at 8:39 PM ^

I'd agree to your point.

However, you have to realize that he's going to be playing in a new scheme. This means not only will he have to adjust to the new mentality of the offense but also 'unlearn' some of what he was taught in the spread offense.

I am sure that denard will be better that last year (in that more comfortable) but the degree of improvement may be limited by the change of scheme.


May 17th, 2011 at 3:38 PM ^

don't go back to the 3 yards and a cloud of dust.  I hate to lose but when you have a kid like D-Rob you gotta let him go.  I think RR knew that and the rest would come around.  Hoke has 3-4 years to be back to ManBall(lol) but only two with this kid.

ND Sux

May 17th, 2011 at 12:31 PM ^

"...Craig Roh as a pass rusher instead of a clunky linebacker ."

was perhaps the most frustrating thing to watch last year.  Field goal kicking gets honorable mention. 


May 17th, 2011 at 12:32 PM ^

These stats are just one of many that give me so much optimism for 2011.  

Everyone tried to point to historical trends in college football and then  apply it to Michigan the last 3 years, but we need to recognize this was just the most cosmic fucked up mess ever imagined.  

The qb situation and the db situation make these last 3 years truly a unique disaster to never be replicated.  

I still stand steadfast that luck/karma/chance will be on our side this year and we will be Big Ten Champs once again.   We just need a return to normalcy not a  MSU luck fest to get it done.

To look at our team last year and consider we still won 7 games thru that mess is an accomplishment.  

My prediction is that we will be Big Ten champs and B Hoke will become the King of Michigan. 






May 17th, 2011 at 12:33 PM ^

I think the most important factor in this equation over the last 2 seasons has been the defense's inability to generate take-aways. Give Michigan average take-aways and the TO margin looks a lot better--I bet.


May 17th, 2011 at 4:24 PM ^


The following table summarizes UM's offensive and defensive numbers (in regards to TO's) under RR:

Year Offense Defense
  Rushing Passing Total Rushing Passing Total
  Fumbles Lost TD's INT's   Fumbles Lost TD's INT's  
2010 29 14 23 15 29 12 7 21 12 19
2009 29 13 15 15 28 14 5 18 11 16
2008 38 18 11 12 30 24 11 19 9 20

UM's offense commited quite a few TO's but the numbers that really stood out to me were the fumbles forced and recovered in 2009 and 2010 by the UM defense. Those numbers are pathetic. I guess when you have trouble even getting to the ball carrier and then tackling him, you can't really worry about trying to strip the ball.


May 17th, 2011 at 12:49 PM ^

...I thought it might be about Turnover (CC) Turnaround (Brian's opinion thereof) Take Three (even though it would have been like take eight). No such luck.

That said, yay regression to the mean.

Now, take a look at this just posted Tom Dienhart article wherein college coaches were asked their opinions on the most important statistics to measure themselves and their opponents. It seems that many of these head coaches and coordinators think that turnovers are incredibly important and that their occurrence is no accident.


May 17th, 2011 at 1:02 PM ^

Accidents aside, that might be the best Dienhart article I've ever seen.  That's not saying much, of course, but it was interesting to see which coaches gave thoughtful responses and which ones spouted cliches.

Much better than reading cliches from Dienhart alone ...

MI Expat NY

May 17th, 2011 at 1:29 PM ^

Most of these coaches are simply stating the fact that turnover margin is the best indicator of whether you will win or lose a particular game.  This is undoubtedly true.  Many of them don't even go into any attempt on their parts to improve that statistic besides "take care of the ball."  I'd also suggest that a positive turnover margin is generally  just an indicator that you have a team that is more talented than your opponents.  A good quarterback means less turnovers.  A good secondary means more interceptions.  A dominating defensive line means more fumbles/interceptions.  

Everything about this team says that turnover margin should get better.  Denard is more experienced (though this may be mitigated by learning a new offense), the defensive front should be better with Roh on the line and Martin hopefully healthy, the secondary should be better as they are more experienced and add Woolfolk back into the mix.  It won't come down to coaching differently except of course that if coaching improves the players, it should also improve the turnover margin.


May 17th, 2011 at 1:44 PM ^

...recognize the importance of turnovers, presumably they're teaching HOW to take care of the ball and HOW to cause fumbles and interceptions. Pat Fitzgerald's quote is instructive:

"The tell-tale one for me is turnover ratio. If you take care of the football and take the ball away from people, you have a chance to be pretty good as a football team. Some would say rush offense or rush defense, red-zone scoring, third downs -- there are a lot of different things to look at. But the first thing that jumps out to me is the turnover ratio. We coach the heck out of it on both sides of the ball."

Every coach talks about how important turnovers are. Perhaps the difference is the degree to which that talk is lip service or something else like a true emphasis on "coaching the heck out of it".

MI Expat NY

May 17th, 2011 at 5:25 PM ^

I just don't think coaching up turnover margin is all that effective.  You can coach a running back or receiver to protect the ball, but you can't really simulate in practice big hits that often cause fumbles.  You can coach a QB on proper reads, but his skill level is probably going to dictate success in the game.  You can coach being in the proper position in pass defense, but whether that leads to an interception is largely on the opposing QB (ridiculous Woodson leaping one handed INT against MSU excluded).  About the only area where I can see coaching having an impact on turnover margin is causing fumbles, and that's only because I've seen the Chicago Bears dedicate themselves over the years to attempting to strip every ball carrier, this also sometimes backfires with poor tackles.  

Turnovers typically come down to good players being in the right place and making plays (the DE that gets a clean shot on a QB,  the CB who expertly reads a route), your opponent making bad plays (QB wildly off target or making horrible reads, punt retun man muffing a kick, etc.) or luck (bouncing ball coming right back to the offensive player, tipped passes falling harmlessly to the ground or right to a defensive player, etc.).  You might be able to coach up some of these particular areas, but in my opinion, you're better served simply coaching your players to be better.   


May 17th, 2011 at 12:36 PM ^

"the turnover margin they enjoy one year has virtually zero predictive value for the turnover margin they will enjoy the next year. That means that on average, teams with substantially positive margins will see major decline in margin the next year, and teams with substantially negative margins will see major improvement the next year."

This is just plain wrong.  If there was no predictive value for year over year turnover margin, then how can we predict major declines or improvement for teams with substantial margins?  That would imply a high negative correlation.  No correlation (or low as is 12%) means that it is random and unpredictable.


May 17th, 2011 at 12:55 PM ^

If you flip 10 coins, and flip 10 coins again, there is no predictive value between the first 10 and the second 10.  r = .00

Now, let's say you got 2 heads and 8 tails on the first set of tosses.  What's your prediction for the second set?  If you said "5 heads and 5 tails," you are of course correct.  So even though there is "no predictive value" from one set of tosses to the next, we can still predict an improvement once we know that the first set of tosses was below average!

El Jeffe

May 17th, 2011 at 1:58 PM ^

I agree with both of you, but you are making slightly different points. drewgrab is saying (essentially) that if you plot turnovers in year 1 on the x axis and turnovers in year 2 on the y axis, r = .12 gets you a big cloud of points with a slight but probably imperceptible upper-left to lower-right quadrant tilt.

This means that teams below the mean in year 1 are equally (more or less) likely to be above or below the mean in year 2. So being below the mean in year 1 doesn't make you more likely to be above the mean in year 2 than remaining below the mean.

Your argument is more about regression to the mean. If you are 119th in TO margin in year 1, you are much more likely to regress to the mean than to stay at 119 or move down to 120.



May 17th, 2011 at 2:26 PM ^

Exactly right--and that's the point being made here.

Teams with a -10 turnover margin can be predicted to be better the next year, even though there is no way to predict where they will be.  Drewgrab said that since there is no correlation, Brian's post is wrong and there is no ability to predict next year.  But of course that's not the point that was being made--the point is that an improvement is much more likely than not.  Nobody (not Brian, not me) was saying that Michigan is more likely to be above average than below average. 

The point was that Michigan is more likely to have a better turnover margin than they did last year, and the reason that this is true is because there is no predictive ability from one year to the next.


Icehole Woody

May 17th, 2011 at 12:39 PM ^

A positive turnover margin means Michigan QBs throw fewer interceptions and do not fumble the ball.  Michigan punt returns do not muff the punt.  Michigan's ball carriers do not put the ball on the ground.

It's something that must be seen to believe.





Other Andrew

May 17th, 2011 at 12:40 PM ^

I can't imagine that a 12% correlation is stastically significant, even with a sample size from here to the moon (which analysis didn't have anyway). Fair to just say that they are not at all correlated year to year.

Additionally, when talking about an overall trend, this kind of correlation analysis makes sense. But when you're talking about just one team and just one pair of seasons, it is irrelevant. Any team could be an outlier at any point. And the reasons Brian mentions are far more powerful data than any look at the overall landscape of college football.


May 17th, 2011 at 12:59 PM ^

Denard threw 11 INT's on 291 attempts in 2010. SDSU QB Ryan Lindley, by contrast, attempted 421 passes

If UM's 2011 offense looks more like SDSU's 2010, one could expect Denard to throw more like 375-400 passes. 

Even if you assume his INT ratio will be lower (not at all a given due to the change in offensive philosophy, and reduced number of incredibly wide open receivers moving from spread offense), it seems reasonable to expect more, not fewer INT's.

Lindley, for example, threw 14, compared to Denard's 11.

On another point, what the statistics Brian references seem to say to me is: from a big picture perspective, TO margin is essentially random year to year. Or at least behaves as though it were random. Of course, coaches still need to coach it, and we all will discuss it. Just don't get too carried away in predicting it. 


May 17th, 2011 at 1:53 PM ^

reduced number of incredibly wide open receivers moving from spread offense)

People have made this argument a few times, but I don't agree with it. How often did we really have wide-open receivers last year? Yeah, there was Roundtree against ND. But it's not like that happened all the time. It only took a handful of games for defenses to catch up to the fairly simplistic routes our receivers ran. By the second half of the season, even the Denard-fake-QB-iso-PA wasn't fooling people. From MSU onward, many of the "wide open" receivers we did have were that way because defenders read Denard's eyes and went to the receiver he was targeting - not because the offense had them fooled. Just the opposite.

Blue in Yarmouth

May 17th, 2011 at 2:01 PM ^

I have tried to block a lot of the past three years out of my memory, but I seem to remember a lot of wide open receivers running down field and DR just wasn't finding them. In comparing it to an offense that utilizes 2 or 3 wr's where you very rarely actually see a wr wide open, this offense seemed to do it on a fairly regular basis from what I can recall from the games I watched. I mean...2 or 3 wr sets would have guys running wide open against our D the last couple years, but not that often against defenses that know how to cover. 

Again,  I may be wrong, but I seem to remember many occasions where I am screaming at my TV for DR to throw to a certain WR running wide open down field only to have him throw into double coverage. 

Promote RichRod

May 17th, 2011 at 2:22 PM ^

Illinois game and another 15 or so plays disagree with you.

There's also another 10 or more plays that got WRs hand-wavingly open and Denard just didn't connect on the pass.  I'd advise you to re-watch the 2010 season (or re-read the UFRs) if you didn't think the spread offense and play designs resulted in a ton of wiiiiide open receivers.

Blue in Yarmouth

May 17th, 2011 at 1:24 PM ^

where I thought I would be craving mediocre....How things have changed for my beloved wolverines.I would take mediocre in a heartbeat now. Let's hope Hoke can bring us back to the glory days.