Gauging Team Effectiveness: Part I: Season In Review
Or, How I Learned to Stop Worrying and Love Hoke
Since it’s the offseason and there’s not much going on, I thought I’d take a look at last season’s team production, offensively and defensively. All the data is from either NCAA.org or ESPN’s drive charts, except for the Air Force and New Mexico games, where I had to extrapolate the drive data from the box score. That’s what’s known as foreshadowing. Offensive and defensive stats are broken out individually, to try to handle those phases of the game on their own.
All Spreadsheeted-up and no place to go
I started out planning to show that the offensive effectiveness remained somewhat consistent through the season, and that only the number of drives per game decreased into the meat of the Big Ten season caused the downturn in scoring. The data did not support that. The number of drives did vary between UConn’s 8 and the Illinois 19-drive trackmeet. But the numbers did not coincide with strength of opponent, final score, or much of anything. You need look no further than the Wisconsin game vs. the OSU game for proof. Against Wisconsin we scored 28 points on 10 drives, compared to 7 points on 12 drives against OSU. The chart doesn’t show any correlation between drives and points:
Chart:
| Opponent | Yards | Drives | Pts. | YPD | PPD | D-Yds | D-Drvs | D-Pts | D-YPD | D-PPD | Net YPD | Net PPD |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| UConn | 473 | 8 | 30 | 59.13 | 3.75 | 343 | 9 | 10 | 38.11 | 1.11 | 21.01 | 2.64 |
| ND | 532 | 16 | 28 | 33.25 | 1.75 | 535 | 17 | 24 | 31.47 | 1.41 | 1.78 | 0.34 |
| UMass | 525 | 10 | 42 | 52.50 | 4.20 | 439 | 11 | 37 | 39.91 | 3.36 | 12.59 | 0.84 |
| BG | 721 | 11 | 65 | 65.55 | 5.91 | 283 | 11 | 21 | 25.73 | 1.91 | 39.82 | 4.00 |
| IU | 574 | 12 | 42 | 47.83 | 3.50 | 568 | 12 | 35 | 47.33 | 2.92 | 0.50 | 0.58 |
| MSU | 377 | 11 | 17 | 34.27 | 1.55 | 536 | 11 | 34 | 48.73 | 3.09 | -14.45 | -1.55 |
| Iowa | 522 | 12 | 28 | 43.50 | 2.33 | 383 | 11 | 38 | 34.82 | 3.45 | 8.68 | -1.12 |
| PSU | 423 | 10 | 31 | 42.30 | 3.10 | 435 | 10 | 41 | 43.50 | 4.10 | -1.20 | -1.00 |
| Illinois*** | 676 | 19 | 67 | 35.58 | 3.53 | 561 | 18 | 65 | 31.17 | 3.61 | 4.41 | -0.08 |
| Purdue | 395 | 15 | 27 | 26.33 | 1.80 | 256 | 15 | 16 | 17.07 | 1.07 | 9.27 | 0.73 |
| Wisc | 442 | 10 | 28 | 44.20 | 2.80 | 558 | 12 | 48 | 46.50 | 4.00 | -2.30 | -1.20 |
| OSU | 351 | 12 | 7 | 29.25 | 0.58 | 478 | 13 | 37 | 36.77 | 2.85 | -7.52 | -2.26 |
| MSU(SEC) | 342 | 11 | 14 | 31.09 | 1.27 | 485 | 11 | 52 | 44.09 | 4.73 | -13.00 | -3.45 |
| averages | 488.69 | 12.08 | 32.77 | 41.91 | 2.77 | 450.77 | 12.38 | 35.23 | 37.32 | 2.89 | 4.58 | -0.12 |
And Graph:
So I need to look a little deeper, namely at typical markers of yard and points. The basic idea is straightforward: good yards per drive equals good “effectiveness” and good points per drive equals good “finishing.” First up: YPD. Offensively, YPD varies from unstoppable against weak competition (UConn, BG, and UMass) to not-very-good against MSU, the other MSU and OSU. Defense, on the other hand, was great in the rain against Purdue, better-than-average against ND and Illinois (per drive, remember), and shelled by MSU, the other MSU and Wisconsin. This should not be news.
Alright, so now we know we couldn’t stop anyone. How about scoring, PPD? Because we’re dealing with a smaller range, I think the data is clearer. Offensively, after throwing out the Bowling Green anomaly (seriously, 6 points per drive?), most of the games turned out be between 2 and 3 PPD, with OSU being a lowly 0.5. Defense tells the rest of the story. The season started out well enough, holding ND to 1.4 PPD, but the number crept up from there, 3 PPD to Indiana, up to 4 PPD to PSU and Wisconsin. Even Tressel-ball managed to score almost 3 PPD. And that chart just looks worse and worse as the season goes on. This is also the point where I get to mention 4-for-14 on field goals and lament.
So what does any of this say? I’d like to be able to adjust some of those values for strength of opponent, so that the 27 points against Iowa’s #7-ranked defense look a little more in line, but I can’t decide on a formula to adjust expected versus actual points. Someone wake up the Mathlete for me, if you don’t mind.
I think that all I can say at this point is that the defense was bad across the board last season, and performed worse against good opposition. The offense was at-best inconsistent, ranging from good to average from game-to-game. With a team consisting of mostly true sophomores at skill positions, I don't that should come as a big surprise.
Coming tomorrow: the same analysis for SDSU's season, and comparisons to what Michigan did.
For what I did to adjust basketball efficiency for an opponent, I used this type of a formula.
(points per drive) / (1-((opponents' points per drive allowed) - (average points per possession allowed in the NCAA)))
Maybe it will work, maybe not.
Methodology explained here: http://maizecoloredglasses.blogspot.com/2011/03/polynomial-centric-big-ten-recap-part-i.html
I'll try to remember to look at that for part II. I think that probably makes more sense than trying to create a scale for the offense/defense rank, when I already have the team's average PPG.
Sorry, I got it wrong.
It's:
A/(1+B-C)
A = points per possession
B = defense's average ppp allowed
C = nat'l average ppp
And I hope you don't mind, but I'm thinking of doing something similar to this for my blog. Just thought I'd ask for permission first, but it's a little different than what you've done here.
Go right ahead, I kind of came up with what I wanted.
Beyond that, the statistical analyses always make me cross eyed.
I looked at all the stats and graphs and I conclude that I hope Michigan wins all their games next season.
nothing will convince me that we would have put up 27 points against Iowa if we weren't already mostly dead in the water in the second half. Maybe that is a closed minded numbers retarded viewpoint. I've watched over 1500 football games since 1982. I can see a defense that is comfortable and coasting. I can point to other league games where the numbers at the end of the game were misleading regarding our offense.
Therein was the central problem that caused it to unravel for the prior coach. The offense was good at best and never consistently good against teams with athletes on defense. Had it been consistently good, with promises of awesome against good teams, I'll bet we'd have a new defensive coordinator, rather than a whole new staff.
But looking forward, there is no doubt that this is an offense with good talent and definite potential. Looking forward to your San Diego State breakdown.
April 24th, 2011 at 11:58 PM ^
Using Iowa as an example:
I agree that a lot of our points came in garbage time against Iowa - although once you get within two scores and you have the ball it's debatable whether or not that is garbage time - but let's do a little run down of the game.
First, let's look at starting field position. On Iowa's six scoring drives, they started from their own 16, their own 49, 48, 37, 40 and 40. So they scored one touchdown on an 84 yard touchdown drive and after that their worst field position was the 37. Other than that first possession, they had great field position. Why is this? Because of turnovers and poor special teams play on Michigan's part. What were our turnovers on these possessions? One was an interception by Denard - the third first year quarterback of Rich Rodriguez's tenure and really was Denard's first year of playing quarterback and not X-factor gimmick - and the other was an interception forced by Tate who was trying to hard to get something going. I'm leaving out the fumble by Smith and Tate's last minute pick because they didn't contribute to the bad field position.
Now look at Michigan's starting field position. We will leave out Michigan's second and third touchdowns because they were down 28-7 and 35-14 when they were scored and thus Iowa probably wasn't going all-out. Michigan started at the 25 and 30 on their first and fourth touchdown drives. Furthermore, on the missed field goal and the Smith fumble - when Iowa's defense still should have been in A-game mode and Michigan should have created points - Michigan started on the 20 and 16. It's also important to note that Michigan's best starting field position was their own 35. Iowa had SEVEN drives start in better field position. They capitalized on this by scoring on 5 of them. So to break it down, Michigan's 4 drives that should have resulted in points were, on average, started at the 23 yard line. Iowa's 6 drives that resulted in points averaged a 38 yard starting position and after the first long drive, they started from the 43.
The point is, Michigan's offense actually performed well when the game was still releveant, but was handicapped by several things:
1. Inexperience at the quarterback position.
2. Drive killing penalties, usually in the form of holding on one RS Freshman Taylor Lewan.
3. Terrible special teams, be it mis-hit kicks out of bounds, missed field goals or a lack of a return game.
4. A lack of short fields created by their defense. Michigan's offense did not get to work from near midfield like Iowa, because they could never force a turnover. Part of this was Iowa's quarterback situation compared to Michigan's quarterback sitiuation. Stanzi was a 5th year senior with a lot of big game experience and this tends to lead to more composure and fewer mistakes (see 1).
In short, the problems with Michigan's offense were all things that were either A. out of the offense's control or B. due to things that you would expect from young players. Only one could really be attributed to an actual problem with the offense: the lack of a solid running back.
I would argue that our offense WAS great at best or at least flashed the potential to be great last year. The only time I felt we were truly shut down and never had a chance was the first half of the Wisconsin game. Most of the problems the actual offense had should have been resolved going into this year. The real question, if Rodriguez was still the coach and even now with Hoke, is how much the special teams and defense could improve to help put our offense in position to succeed to the full potential.
Of course now the questions are all about Al Borges and what he decides to do with our team's greatest strength.
April 25th, 2011 at 12:27 AM ^
but not the conclusion you draw from it.
Yes Denard made a lot of turnovers and was a first year starter. IF he doesn't make those mistakes then the offense would have succeeded more.
But, you can't say, "Michigan had an awesome offense when you take away the drive killing turnovers and assume that TD's would have been scored". What you really have to admit is that Michigan's offense was just average to good against the tougher defenses in the conference.
Maybe the mistakes are corrected in the following season, and maybe they aren't, but it really isn't a logical argument to state an offense is awesome and consistent when the results are not awesome and consistent.
I was very excited about Denard after the ND game also. I was excited in fact all the way up to his first second interception in the MSU game.
I think if you take the Illinois game out of FEI then Michigan is not ranked #2. That game had crazy stats, and got me excited, but the performance could not be repeated.
So you can't say that the entire season was an awesome offensive performance, because it just wasn't. Points per drive against OSU demonstrates that clearly. So does the bowl game.
I don't see how you could take that from my post.
My point wasn't so much to argue that the offense was whatever adjective you choose - average, good, great, awesome, TREMENDOUS - but rather they were handicapped by a variety of things that prevented them from succeeding at a higher level. For instance, if we had a defense that finished 69th overall instead of 109th and a field goal kicker that could hit 3/4 of his kicks inside the 40, I would garauntee our offensive performance against "tougher" defenses would be far superior than it ended up at the end of the year. Part of what makes each unit is how the other two perform.
I also like that you brought up the Ohio State game. It's another great example of what I laid out above regarding Iowa. Ohio State took a 17-7 lead and... OSU's drives consisted of two 3 and outs, a field goal in which they received a holding penalty that prevented a first and goal - because longer drives have more chances to result in more mistakes - a 35-yard touchdown drive that came about from an 18-yard punt and a kick-off return for a touchdown. That's the difference. If you were comparing Michigan and OSU's offense at that point in the game, which was the better squad? However, OSU scored 17 points on the back of superior defense and special teams.
Finally, while I know you can't take away certain elements of a performance and call it better than it is - i.e. Denard's turnovers - I was commenting on the fact that over time, those things should improve, because as players mature, they tend to cut down on their mistakes. It's just the way things work. It doesn't mean that last year we had a consistently awesome offense, but it bodes well for 2011-12 that many of the mistakes our offense made were things that should be improved upon simply from more experience.
In the same breathe, you can't just cut out certain games from last year and use those hypothetical numbers instead!
"I was commenting on the fact that over time, those things should improve, because as players mature, they tend to cut down on their mistakes."
I wish this were true, but haven't we been hearing that exact same line about 1) turnovers and regression to the mean, 2) defense (it can't get any worse next year), and 3) field goal kicking (it can't get any worse, they're getting older) for two years now. Prior to 2009 I would have believed your assertion that things simply must get better with time by definition. Hopefully that will be the case this year with some of those items, but I've learned not to just assume improvement must occur because of experience.
April 25th, 2011 at 10:34 AM ^
and what I am saying.
Who committed many of the backbreaking turnovers? Denard. I've already explained his situation.
Who committed many of the silly penalities? Lewan, I've already explained his situation.
Our kicking game? It wasn't bad in 2008 and 2009. It was only last year, when Gibbons - First season as a college kicker! - when it fell through the basement. (admittedly, it doesn't look as if he is getting better, but perhaps another season will feature his improvement. Hopefully Wile is at least competent).
Our terrible defense? While you have a point on the linebacking corps, our secondary was once again full of youth and players playing out of position.
Once again, it's not everything - Vincent Smith seems to be a fumble machine and there's no explanation for things like 87 dropped balls against Ohio State - but a lot of Michigan's woes came from youth and inexperience. Things that can really only be cured with time.
First, let's look at starting field position. On Iowa's six scoring drives, they started from their own 16, their own 49, 48, 37, 40 and 40. So they scored one touchdown on an 84 yard touchdown drive and after that their worst field position was the 37.
While all this is factually true, did you have any confidence that day that Iowa wouldn't have scored on those possessions if they'd simply started with worse field position? They pretty much did what they wanted to on offense, and only the endzone seemed to keep their yardage totals down.
That's not really my point. Although I do believe a longer field allows for more opportunities for things to go wrong.
For what it's worth - and I don't think its much - on drives started inside their 30, Iowa stalled on 3 of 4.
My question back at you is, if Michigan had Iowa's field position, do you think they would have scored more relevant points than the ones I listed? Michigan's offense was not put in positions to succeed. Iowa's defense/special teams - or our special teams - gave the Hawkeyes' offense a better chance at success.
There is no argument that Iowa's offense couldn't score on Michigan's defense. Michigan's defense was terrible last year. That is a fact.
What is in question is the quality of the offense and how much their lack of success/ success - depending on how you look at it - had to do with things within their control or out of their control.
Looking at the drive chart http://espn.go.com/ncf/drivechart?gameId=302890130, we scored on the opening drive of 75 yards. Only one drive in the first half was three-and-out. There was a missed field goal and an INT at midfield that set up an Iowa TD.
I won't argue that there were ball control problems, but one came in the red zone, and one came from Tate in the 3rd after Denard went out. Iowa benefitted from short fields, which the net YPD shows.
The Wisconsin game we got shut out in the first half, but I think the Iowa game is in the "what if" category.
Regardless of your conclusions, I vote yea for the Dr. Strangelove allusion.
"You can't fight in here! This is the war room!"
Seriously, thanks for the work and I look forward to part 2.
I assure you, any Kubrick allusions were accidental.
Oh right, the Dr. Strangelove section title. Sometimes I'm too smart for my own good.
Love "the" Hoke. Just sounds funnier.
April 24th, 2011 at 11:35 PM ^
I like what you posted but I wish I could read it all
Thank you.
April 25th, 2011 at 11:20 PM ^
Our offense wasn't even that good, in terms of PPD. Only 4 games above 3 ppd??
Wisconsin, a real Big 10 contender, averaged--averaged--5.882 ppd.
http://www.seldomusedreserve.com/?p=98
Net PPD is even worse for us, as it takes into account both our terrible defense AND our prolific, yet inefficient offense...
Comments