I would think that having a brand new would be a variable that could contaminate predictions. Just something to consider when lumping UConn in there. But yes, I am expecting an epic game that may end up foreshadowing the season.
Can Michigan Expect Another Major Improvement on Offense in 2010?
[Ed: meant to bump this sooner but there was a lot of stuff yesterday.]
After the disastrous offensive performance of 2008, the 2009 Wolverine offense really had nowhere to go but up. Using my offensive ratings, the 2008 Michigan offense was 7.4 points per game below average, 107th out of 120 FBS teams. 2009 brought another year in the system and real quarterbacks and huge improvements. While far from consistently excellent, Michigan moved up to a modest 1.2 points per game above average, 50th nationally. No one outside of the eternal optimists like Fred Jackson could see another 57 place ranking improvement, but what has happened to teams that have shown big offensive improvements in year in the following year.
Presently my database has the 2007-2009 years completed, just enough for a 3 year case study. From 2007 to 2008 there were 28 teams that improved offensively by at least 5 points per game. I broke those team into three categories, teams that saw a second major (+5) increase in the third year, teams that saw a major (-5) regression back in the third year and teams that were in the middle and didn’t necessarily continue gaining, but didn’t fall back much either.
*Only BCS teams shown
With 14 of the 28 teams in this group, half of the teams that show big gains can expect a return to the mean the next season. In fact, these teams were worse offensively in 2009 than they were in 2007, let alone the beacon season of 2008. The average team in this group was 2.5 points per game worse in 2009 than they were in 2007 before they peaked.
The closest thing to a consistent thread is the quarterback possession as five of the eight, Oklahoma, Baylor, USC, Arizona and Utah, spent most or all of the season with a new quarterback.
In general, the regressers look like a group that is just regressing to the mean and that replacing a quarterback is damaging when your success has not been sustained for longer than a single season.
With the exception of Alabama, these teams were pretty average in returning starts and had no major position group gaps to fill. Alabama had a new quarterback and was 97th in returning offensive starts nationally, the ability to sustain the offensive success is likely attributable to the influx of talent Saban brought into Alabama since he arrived.
*Michigan 2007 results omitted (-1.1)
With a relatively new coach and a total offensive system overhaul, Georgia Tech is clearly the most similar situation to Michigan and their path is one that Michigan would be thrilled to follow. Tech went from –1.1 ppg in 2007 to 7.6 in 2008 to 14.5 and my top rated offense in the country in 2009. Even though Johnson and Rodriguez were hired the same year, the Michigan offense is about 2 years behind Georgia Tech. Georgia Tech went from average to very good to best in the country. Michigan went from average, to very bad and back to average. Even with the offset timeline, Michigan seems comparable to Georgia Tech’s situation and therefore a second year of offensive gain seems very possible under this comparison.
All five of these teams either returned 20+ starts at the quarterback position (except GT who had the same quarterback from the start of the system), although Stanford’s returning quarterback was replaced. The other major similarity between these schools in neither of the last two years did they have stratospheric gains, there is less flukiness to these teams success.
When looking at the progression from very bad to roughly average, there are four BCS level schools who showed that same progression. Three of those (TCU, Notre Dame and Pittsburgh went on to see big gains in year 3 as well, and NC St still saw modest improvement. Teams fitting this profile for a potential second year of strong offensive progress in 2010 along with Michigan include Kentucky, UConn, Wake Forest and Mississippi St.
Although teams that show a big jump like Michigan last year are more likely to fall back than continue the progress, the recruiting profile, experience at quarterback (even if the returner loses his job), progressions comps and system change all point to Michigan as being a good candidate to at least sustain and probably show more improvement next year. Every 3 point gain is worth about one additional win on the season and based on this look I would say that from the offense alone, a 3 point gain seems likely and a 6 point gain entirely possible.
Ummm UConn doesn't have a new coach...
Tate should be much healthier.
Denard should be much better.
The O-line should be much more solid (Molk) and settled.
The tight ends are the same or better.
The wideouts have a lot more experience and a year's growth behind them (as well as an expanded and more internally competitive group).
The "running back position" can't be any worse than last year (due to unfortunate strings of injuries to guys like Minor and Brown who worked their butts off but were hobbled much of theyear).
One more year of overall team development with RR and Magee.
How can all of that not translate into "improvement"?
if the defense is any better (and we can all agree, they have to be) then there will be more opportunities for the offense, better field position and the other team's D gets tired out from being on the field so much. So yeah, I'm expecting continued improvement from the O. In fact, you ain't seen nothing yet!
Agree.I thought I heard in one of RRods pressers that the QBs were gonna get the whole playbook.
Mathlete's analysis is comforting, for sure. But more comforting, the squad only lost five players (at four positions) and none did better than signing post-draft flyer contracts. Of those, I personally don't see any sticking with their teams except maybe Minor because he's a good, underrated RB and Mathews because the Bears need WRs badly.
Depth, experience, continuity and Barwis are our friends.
It's good to have Barwis on our side. If I ever anger him, may the Super Friends help me.
I haven't been following your posts closely, but how do your offensive ratings deal with the Delaware St. game? How does your analysis change if Baby Seal U is removed from Michigan's statistics?
My math shows that Michigan averages over 3 fewer points per game if you remove this game from the schedule.
If you remove Delaware St, then you need to remove everyone's biggest cupcake--which would probably result in very little relative changes.
Everyone plays shitty teams.
Everyone plays shitty teams, baby seals, cupcakes, your team, my team, whatever.
On the same note, I wonder how the analysis changes if we remove the Penn State game from Michigan's statistics.
Actually, I don't, because it doesn't matter. The team plays 12 games, they all count. EVERY ONE of the teams on this chart plays at least one soup can, including Florida, Alabama, Texas and Oklahoma.
The point is moot because Delaware St. is not included in Mathlete's calculations.
Still, I disagree with you strongly. Everyone does not hang 63 points on their worst opponent. For example, Michigan State (44), Ohio State (45), and Penn State (52) all scored fewer points in games against poor opponents, suggesting that our points-per-game average is biased upward. Basically, I think most BCS teams could have also hung 63 on Baby Seal U, so, as a data point, it is pretty irrelevant.
I think that looking yards per play would be a better metric, anyway.
Let's also remove all plays where it didn't look like the other team was trying very hard.
JK - sometimes you have to laugh
All games versus FBS opponents are always included in my analysis and no games versus FCS opponents are ever included.
I think that your Delaware State comment is legit. Michigan played seven common opponents between 2008 and 2009. ND, MSU, PSU, Ill, Purdue, Wisky and OSU. They scored 151 against these seven teams in both seasons. That accounts for 58% of the schedule in each season so it should be a fairly representative sample.
What's even a little more telling is that Michigan scored FEWER points against five of the seven teams in 2009. The +21 differential against ND made a big difference.
While we could all visually tell that the offense was a little better in 2009 (especially early in the season), I think that gets blown out of proportion at times.
Hopefully the Molk injury -- which happened before most of those games -- is a mitigating factor in the slide in offensive production as the season progressed.
So there would be no Delaware St. bias.
because he said it two comments ago.
I know you were looking for more diary topics. Another similar one that you could do would be defensive efficiency from year one to year two in a system. You could also look at efficiency vs. returning starters. It would be interesting to see what to expect out of our defense next year and it would go hand-in-hand with MCaliber's diary where he speculates the number of wins based on the offensive and defensive YPG.
This analysis is just awesome. Are you planning on doing a similar analysis of the defense (perhaps an analysis that makes us all a bit les scared about the defense this season)?
I may be in the minority here but I don't think any kind of statistical analysis of 2008/2009 is relevant for 2010.
2008 is a complete throw away year. We all know that QB play is 90% of RR's system and 2008's two QBs are no longer on the Michigan roster. Therefore 2008 stats are completely meaningless and there is almost no carry-over from that season to 2010 except for a few OL and a WR or two.
2009 has some relevancy for 2010 but the way things are going (Tate not improving apparently and DR improving greatly) I just don't think 2009 is all that relevant to what we should see this fall.
If anything year 3 should improve over the past from an intangibles standpoint. Returning players understand the system better, coaches have a better understanding of what works against which defenses, etc.
Your statistical analysis is great and relevant generally but I think in this case it's not a useful prediction tool for 2010's offensive success.
I was thinking about the loss of Brown and Minor the other day and wanted to get a sense of what we had coming back in the ground game and then got carried away and looked up the numbers for every phase of the offense. So here are the rough totals for what we have returning as far as career stats go:
Rushing: 1,400+ yards and 15 TD
Receiving: 2,500+ yards and 18 TD
Passing: 2,200+ yards and 15 TD
You can also throw in two return game touchdowns from Stonum/Odoms.
Seems like a decent amount of production but I wonder where it would fall nationally and if that means anything as a predictor of future success. The best part going forward is that none of that production graduates this year except for Moundros and Webb, so barring early departures the 2011 offense should be loaded with returning talent/production.
I seem to remember some talk on this blog last year about a Wall Street Journal (I think) article that showed how returning starts on the OFFENSIVE LINE had a direct correlation with number of wins. Of course, last year Michigan was highly ranked nationally in terms of returning starts at O-line, so that must not always work out, but the law of averages would lead one to conclude that another year with quite a few starts on the offensive line will likely result in a good number of wins if the correlation is true.
Plus, we lost Molk in gave 4. Arches need keystones.
I expect us to jump up in the top 25 for total offense this year. I just feel we have the talent for that to happen.
The conventional wisdom in college football is that teams regress, not to the mean, but to their historical norm—whatever that is.
If Northwestern leads the Big Ten in offense next year, it’s an improvement far less likely to be sustained than if Michigan does it, simply because Northwestern is historically below average, and Michigan is not.
One could give endless examples. Florida State and Michigan State were both mediocre last year. If you had to bet on a team getting better next year, you’d pick the Seminoles, because FSU has a long history of sustained excellence (albeit not the last few seasons), and the Spartans do not.
So it would be interesting to see the patterns when you compare teams, not to the NCAA Mean, but their own mean. An average year for Michigan and an average year for Michigan State are very different things.
Ok, I'll accept the premise, but their mean is likely around .500, as you would expect. A quick calculation shows that Michigan's historical record, best in the country, is 8-4 or 9-3. (.7366 * 12 games). After about the top 25, everyone is looking at around 7-5. So, essentially, regression to the mean most likely is 6-6.
The numbers get a little different if you take, say, the last 30 years instead of all historical, but the point is still the same - that the best winning percentage is >.80, which is still 9-3. So I think it's basically nitpicking to try to differentiate between historical winning percentage, and traditional mean.
I am still working out the kinks, but my general approach is similar to regressing to the mean, but instead of regressing to a historical average - which would make Miami Ohio and Central Michigan top 20 programs and Oregon out of the top 50 - I am regressing to a weighted recruiting history. The better your recruiting over the last 4 years, the better your mean is. The outcomes are not all that dissimilar, but the distinction is important.
I'm not sure if it'll be a major improvement but I'd like to think there will be a pretty nice improvement.
UM's offense was much better in the non-conference games in 2009 then in 2008 even when you exclude Baby Seal U. However, in 2009 UM ranked in the lower half of the big 10, if not very close to the bottom of the big 10, in just about all offensive statistical catagories. I'd like to think that UM's offense will be improved enough so that UM's offense will statistical be in the top half of some offensive catagories in 2010, and not close to the bottom in any.
will help too.
putting a handle on it?
National Rankings (Michigan offense):
2009 - 41st and 29.5 ppg
2008 - 101st and 20.25 ppg
National Rankings (Michigan defense)
2009 - 72nd and 27.5 ppg
2008 - 84th and 28.92 ppg in 2008
The defense was actually slightly better from a points allowed standpoint in 2009. That was not what I expected. But before you get too excited...
Big Ten Rankings (Michigan offense):
2009 - 9th and 22.1 ppg (-7.4 ppg from total 2009)
2008 - DFL (11th) and 20.2 ppg
Big Ten Rankings (Michigan defense):
2009 - DFL and 33.2 ppg
2008 - 10th and 28.9 ppg
I put more stock in the Big 11 numbers just because it is closer to an apples-to-apples comparison year over year. By that measure, from a scoring/stats perspective, 2009 was really no better than 2008 if you believe that "better" means not in the bottom 3 in the league in both offense and defense. Of course, we also had 2 B11 wins in 2008 versus only 1 in 2009.
In terms of expectations, it is very hard to believe that we could go from essentially last in the conference two years running (in total) to anything higher than 6th or 7th in 2010.
An extra two wins puts us at 8th. An extra three wins puts us at 7th. An extra four wins puts us at 5th. That assumes that we didn't beat the teams we were tied with. If we beat Wisconsin, PSU, and Iowa (I know its highly unlikely, but it makes the point) in those 4 games, we are now in the lead for 2nd after tiebreakers.
Realistically, let's look at the past season with more consistency on offense and a defense that allows slightly fewer PPG. We could have easily beaten MSU, Iowa, Illinois, and Purdue. There are your four teams. That would have placed us 4th in the conference.
I don't think it is unrealistic to expect that we end up with 8-9 wins this season and in the top half of the Big10. Especially with the reports coming out of spring practices that most of the powers we will be facing have had troubles with their offenses in their spring games. I am not trying to predict the outcome of the season here, but I think it is a little discrediting to the team to say that those assertions are unrealistic.
...may be an optimist, but I don't even think he foresees another 57-spot jump in the rankings from our 50th place finish last year.
What an interesting observation. What is this, "humor"?
With the eye of Sauron, (NCAA) firmly on RR and Michigan, my only concern is how much extra practice time we're gonna miss from last year!
Great analysis. Keep up the great work.
Michigan is still a very young team on both sides of the ball. By year 3 of the new football program, offensive system, workout regimen, practice regimen, etc., one would expect a much more smooth offensive showing. Their were bright moments last fall, but it was still choppy with injuries and inconsistent play. I would expect things should be more consistent, sure and smooth in 2010 at QB, WR and OL. Michigan is going to continue to run the ball well, I think. But godamn, Michigan remains one of the youngest teams in the country.
another huge reason why huge improvement is possible - turnovers. If these guys can stop killing drives by not hanging onto the ball, they should improve by a few points a game on average.
Having a player with talent who is running the system for the second year will provide the biggest lift. I'm worried as hell about the D, but I expect the offense to be closer to the six points better this year.
RR hasn't been here 3 years
UM 2008 offense was historically awful. No matter what is said about attrition and talent level, there was enough there NOT to lead the nation in 3 and outs. Choosing Sheridan and McGuffie to lead the offense and putting the square peg Threet into the round hold offense was...uh...stupid. But that's over.
Now, in 2009, we did improve. If you look a little closer, against teams we played that had a pulse (ie big ten conference), we actually scored about the same number of PPG. Thank Delaware HS for beefing up the stats.
In 2010, I expect a "real" improvement on offense. We've consistently outrecruited every team in the big ten (sans OSU, and it's close) and have plenty of returning players. This will be an exciting team to watch for sure.