Michigan offense actually better against stronger defenses
So, I keep hearing how the offense put up gawdy numbers against bad competition, but disappeared against good defenses and wondered if that was actually accurate. I took my numbers from my season long diary tracking yards gained by UMs offense.
With a small sample size (51+(5), 50-26(4), 25-(3)), it's tough for these numbers to be something majorly concrete, but it's interesting to see that UMs offense was more efficient against better defenses.
November 28th, 2010 at 7:48 PM ^
But yet we lost by more points against these teams. So who cares.
November 28th, 2010 at 7:52 PM ^
Well, sure, youth on offense will do that. That is exactly why I took yards gained and not offensive scoring. Using yards gained shows an offenses abiltiy to move the ball regardless of age, turnovers, etc...
November 28th, 2010 at 7:54 PM ^
about football is, though, that "offensive ability" is not independent of little things like "turnovers" and "age" - it's about putting points on the board. A young team should have more variability (sometimes losing to crap teams, sometimes beating good ones) instead of being consistently outclassed by good teams and marginally defeating bad ones.
November 28th, 2010 at 8:01 PM ^
I like how you equate offensive success to team success there. Didn't we put more points on wisconsin in a half than Michigan's put on them in a decade? Points or no points a young team will win when it has more points than the other team. Just a little football protip for you there.
November 29th, 2010 at 12:16 AM ^
we score in the 1st half when it was still a game?
Our offense has tons of potential, but I dont care how many yards we racked up..we consistantly could not score against good teams.
Against MSU we blamed it on unlucky int's Against Iowa it was fumbles and int's. Against PSU more or the same. Wisconsine we did absolutly nothing in the 1st half. Against OSU it was fumbles. Bottom line we havent scored mre then 10 points in the 1st half in any of those games..I dont care if we score when were down by more then 3 touchdowns.. it's meaningless. Against EVERY physical team we played we went down like a 2 dollar whore. Nothing you say can change that.
November 28th, 2010 at 10:31 PM ^
Isn't the sign of a good offense points scored? Moving the ball means nothing if your not in the endzone.
November 29th, 2010 at 8:20 AM ^
I'm not getting the rash of "it doesn't matter" replies. Either the only statistic that matters is the W-L column or we can factor other knowledge, like that provided here, when gauging how good or bad a team is or how much progress has been achieved or whether the offense has "disappeared" when facing better defenses. All this "but they put up less points" junk is anti-intellectual and almost beside the point. We know. 'Kay? Moving the ball does mean something, if it's not the most important thing.
November 28th, 2010 at 8:05 PM ^
That is exactly why I took yards gained and not offensive scoring. Using yards gained shows an offenses abiltiy to move the ball regardless of age, turnovers, etc...
I don't follow. Youth affects scoring but not yardage? And isn't the goal of the game to outscore the other team?
November 28th, 2010 at 9:29 PM ^
.......but little things like the youthful mistakes killing drives prevents points from being scored. At times, against every team we've played this year, our offense moved the ball with ease, including against some very highly rated defenses. Looking back at the things that stopped many of those drives from scoring points, the mistakes seen from young players were more prominent than good defensive plays from the opposition. Case in point, the two first half drives against MSU that ended in endzone picks. Those throws were more a case of Denard forcing a throw to a wide open WR than the defensive player making a great play.
November 28th, 2010 at 9:37 PM ^
Inexperience = variability.
Variability is more likely to effect scoring than it is total offense.
Why? Because, in general, you need to string several plays together in order to score (unless you're Denard). You can have 5 positive plays in a row and then one negative play (turnover, missed field goal, etc.) and you don't score.
This pretty much sums up Michigan's season.
November 28th, 2010 at 11:30 PM ^
Keep believing, guys. The excuses have become more sophisticated through time.
November 29th, 2010 at 7:24 AM ^
Not sophisticated at all. Just logic Jack. Good old fashioned logic.
November 28th, 2010 at 8:00 PM ^
So, the numbers unsurprisingly show that we had less yards against better defenses. What is the big surprise? That the ratio of our yards against our opponents average is higher for better defenses than worse defenses? I think you're really stretching. Might want to step away from the stats for a second.
In any case, those numbers are certainly skewed by our performance against Illinois, and outlier even before you factor in the extra 75 yards we got in overtime.
November 28th, 2010 at 8:23 PM ^
Illinois is actually the #40 defense and the OT yards are not factored in. The top 3 defenses were OSU, Wisci, and Iowa.
November 28th, 2010 at 8:37 PM ^
It's kind of hard to say that our offense was "better" against those three teams than against the rest when we averaged 21 ppg in those three games and 38.8 in the other nine games.
November 28th, 2010 at 9:37 PM ^
Sure, but those top 3 defenses are only giving up 16.4 points per game. The other nine are giving up 24.25 pts/game.
November 28th, 2010 at 9:42 PM ^
So against bottom 50 teams we scored 14.6 points per game more than our opponents' average, and against top 25 teams we scored 4.6 points per game more than our opponents' average. Doesn't look as great if you look at numbers rather than percentages.
This is a clear case of cherrypicking stats in an effort to prove a preconceived point. It's silly.
November 28th, 2010 at 9:47 PM ^
That's why they are percentages. Scoring 60 against a Bowling Green team that gives up 30 points a game is similar to scoring 28 against a Wisconsin team that gives up 14. Points are harder to come by against these teams. Pecentages are neat like that.
November 28th, 2010 at 9:59 PM ^
Actually, no its not. Scoring 30 points more than Bowling Green's average is not equivalent of scoring 14 points more than Wisconsin's. Which is why no one other than the OP has used this metric (percentage of points scored versus respective opponents average) in assessing a team's performance in this way.
I understand the concept of "percentages", but the percentages OP is using simply don't prove anything meaningful here.
November 28th, 2010 at 10:02 PM ^
Apparently you do not understand percentages, regardless of what you say. Scoring on Wisconsin is more difficult than scoring on Bowling Green, and if we're looking at exceeding averages, percentages are a good way to do it.
November 28th, 2010 at 10:49 PM ^
You're right, you understand percentages better than I do. The fact that we scored 4.6 more points per game than opponents' average against top 25 defense, and 14.6 more points per game than opponents' average against other defenses, proves, in OP's word, that our "offense is actually better against stronger defenses" than weaker ones. That I think these metrics (which no one else in the history of sports statistics have ever utilized) are not helpful shows that I don't understand percentages, fractions, or long divisions.
This is a half-assed (ie, 50% of the ass) analysis.
November 29th, 2010 at 9:46 AM ^
Agree that OP's title is a stretch. But his math is sound. And it does tell us about the potential of this offense and it gives insight into what's wrong with it.
Disagree on your version of %'s. I'm assuming you grasp this, but the base is completely different for the 4.6 pts vs. 14.6 pts.
The problem I have with a straight pts. analysis is that it tells you WHAT the problem is but it tells you nothing of the WHY. That's where these "crazy metrics" come in handy. Some of us want to understand WHY before reacting to the WHAT...
November 29th, 2010 at 9:51 AM ^
...we averaged 6.8 points in the first half of our games against MSU, Iowa, PSU, Wisconsin and OSU. It's a lot easier to put up yards and points when the opposition is trying to not give up a big play.
November 29th, 2010 at 10:22 AM ^
I don't think you can really say MSU was in a prevent defense when UM was gaining their yards. Denard and company were moving the ball quite well the first half. UM gained 263 yards the first half of that game and only came away with 10 points.
If Iowa was in prevent and not trying to give up a big play, they sure did a bad job of that. UM had 5 plays of 15 yards or more in the second half. UM still gained 200 yards of offense in the first half and had an INT and a missed FG. If you ask me, I just think UMs playcalling became more aggressive.
Same with OSU. UM gained 235 yards of offense in the first half on a defense that averaged 241 yards given up per game.
The Wisconsin game, yeah, that was an abomination.
November 29th, 2010 at 10:49 AM ^
What you are pointing out is not without merit, but as has been mentioned before the objective is to put points on the board. It is sad that we are hanging our hats on the fact that we gained better than average yards against good defenses. We didn't win the games or even come close losing by an average of 17 points in those five games. Points and when they are scored are much more important than yards gained.
November 28th, 2010 at 9:16 PM ^
Didn't say they were top 25. That just shows the numbers for the 50-25 defenses are skewed -- taking Illinois out of the equation, the ratio of our yards versus opponent's average is higher for bottom 50 teams than for teams 50-25, which undermines your point.
In any case, the three groupings you chose are arbitrary (why not below 60, 40-60, 20-40, etc), as is the very idea of judging offensive performance by ratio of yards compared to opponents' average yards given up. I just don't think this proves anything.
November 28th, 2010 at 9:35 PM ^
So, you want me take take part of the equation out because UM had a strong day offensively? Should I also take the Iowa game out too? I might as well remove the Purdue game too because the rain effected both offenses.
Actually, if you take away UMs effect on Illinois' average they are at 309 yds/game, and would be the 14th ranked defense, so the top 25 section get skewed even more towards this offense being better against top defenses.
That change looks like this
November 28th, 2010 at 9:38 PM ^
Well, yes, outliers are typically removed from the equation in statistical analyses. Sure, take Purdue out too. In any case, these numbers are just not statistically relevant, and even if they were don't prove anything.
November 28th, 2010 at 11:02 PM ^
Outliers are commonly removed from statistical analysis but not typically removed. There is no widely accepted method to remove outliers. Usually they are so far outside the rest of the data that it is considered an error. To remove them just because they are outliers with such a small sample size is, what The Wire taught me is called, juking the stats.
I think the Illinois game is the only consideration in the extreme value vs outlier argument but it's probably extreme value given our offenses propensity to put up yards on anybody.
November 28th, 2010 at 11:20 PM ^
The Illinois game isn't as much of an outlier as you are suggesting. UMs offense is gaining, on average for the season, at 147% of their opponents defense. The lllinois game is at 194%, while the MSU game is at 112%. Taking the "outliers" out, UM is still gaining at 154% on the top 25 group, 7% above the season average.
November 28th, 2010 at 8:03 PM ^
Where are those straws? I can stretch my hand out and almost reach them............
November 28th, 2010 at 8:10 PM ^
You should took at turnovers against top teams, becuase that is what killed us all year.
November 28th, 2010 at 8:10 PM ^
I have an idea. Let's trade our defense for an actual one. Kidding! No, not really but really...
November 28th, 2010 at 8:16 PM ^
What your record says you are.
-Bill Parcells
November 28th, 2010 at 8:46 PM ^
“No matter how much you've won, no matter how many games, no matter how many championships, no matter how many Super Bowls, you're not winning now, so you stink.”
- Duane Charles Parcells
November 29th, 2010 at 8:25 AM ^
Charlie Weis said that also - sure he got it from Bill. Unfortunately it didn't come with HC skills.
November 28th, 2010 at 8:16 PM ^
Some empirical evidence. Kudos.
I'd hate to see the similar figs for the defense...
November 28th, 2010 at 8:19 PM ^
for worst. post. ever.
If this were remotely true, then we would have been winning against OSU after the first quarter.
If this were remotely true, we wouldn't have had 3 turnovers yesterday.
If this were remotely true, then even with our horrible defense no team would have gained greater than a 7 point lead on us at any given time this season.
If this were remotely true, we would have won at least one of those games against the good defenses, in spite of the putrid defense. Illinois doesn't count, their defense was exposed by us and was more "good" than "great" overall, and were just overrated at the point at which we played them.
I don't mean to be an asshole (probably too late tho), but people on this blog need to stop with the affinity towards stats. 90% of them are made up. And the other 10% contains many ways for people to pick and choose what they want to argue. While it was encouraging to see teh offense move at will in the first half yesterday, ultimately their demise was a distinct inability to score points in the red zone. That is a major issue. It's a good offense, but you can't exactly call them world beaters quite yet...just too inconsistent this season.
November 28th, 2010 at 8:27 PM ^
I forgot the most important thing I wanted to say! Ultimately, we need to step back and look at the product. Stats are stats, each one means something but can only get you so far. You cannot cherry pick one and say with definitiveness what's "good" and "bad" about Michigan. You have to look at everything as a whole. And in general, if it doesn't pass the eyes test, then it's just not true (which it definitely doesn't...Michigan has looked much shakier on offense against the best defenses than the worse defenses, with the exception of Purdue but only because it was a rainy slopfest on real grass field)
November 28th, 2010 at 8:51 PM ^
Well, yeah, because you know, good defense are good at stopping good offenses. This is why it was broken down on an average and shown that over the course of the season, UMs offense has gained more yards and scored more points on the best defenses than they typically give up in a game. If UM can develop a good defense, this team could be sick.
November 28th, 2010 at 9:35 PM ^
you missed an important fact: your post doesn't agree with what he thinks, therefore it's the worst post ever.
hope that cleared up some confusion.
November 28th, 2010 at 10:55 PM ^
But thank you for putting words in my mouth.
If you're going to evaluate the offense in such a manner, you need to look at everything, not just yards. You need to look at points scored, turnovers, drops, average starting field position...etc. etc. etc. The stats from the OP show that yes, Michigan has the ability to move at will against any defense. But that doesn't mean they've done any better against the best defenses. What defines "better" anyway? The bottom line is the offense scored 7 points against a good defense, and while they looked good at times ultimately they failed when it matters the most.
And I think there is value in stepping back and actually evaluating the product on field in a subjective way rather than stats stats stats.
Am I insane in thinking that way? Apparently, judging by all the negs I'm getting, I must be.
November 28th, 2010 at 11:26 PM ^
Well, I have yards and scoring done for you, I look forward to your analysis on turnovers, drops, starting position, special teams, etc, etc, etc. I'm so pumped.
November 28th, 2010 at 8:59 PM ^
...Michigan has looked much shakier on offense against the best defenses than the worse defenses...
Wow, please tell me more. This statement can be made about every offense in the history of football; literally and without hyperbole.
November 28th, 2010 at 8:20 PM ^
Our O averaged 20 PPG against OSU, MSU, Wiscy and Iowa. They had a total of 7 points against OSU and none in the second half. Iowa and Wiscy controlled us in the first half.
If the numbers make you feel good, that's great, but enough is enough. I care about one statistic - W's and L's. Nothing else matters.
November 28th, 2010 at 8:21 PM ^
Why come he posts something positive and gets yelled at? Why come?
November 28th, 2010 at 9:37 PM ^
why come... you don't have a tattoo?
November 28th, 2010 at 9:54 PM ^
Idiocracy was a great movie
November 28th, 2010 at 8:27 PM ^
I get it guys; you want Harbaugh. This is a perfectly understandable and reasonable desire, but insisting that the offense is not that good is just laughable. Furthermore, refusing to believe that youth has anything to do with turnovers and the lack of scoring is laughable. By all means, root for Harbaugh, but do it for the right reasons, and stop fooling yourselves.
November 28th, 2010 at 8:33 PM ^
Who said that? I didn't. I never said the offense sucked. But we have to be honest about the things we say here. And when someone posts ONE stat and says "our offense is better against good defenses than bad defense", well I'm sorry but it's just not true.
Has nothing to do with Harbaugh. In fact, I think (tentatively) RR deserves the 4th year, because the offense has made such great strides. But we have to be honest with ourselves in the evaluation of the offense too. They've struggled against good defenses.