The Greatest Wisconsin Scoring Debate: Whom do the stats support?

Submitted by Fuzzy Dunlop on

This issue has been discussed ad nauseum.  Certain posters, myself included, think that our offense performance against Wisconsin was not as impressive as the numbers indicate, since we were shut out in the first half and only put up points when we were already down by 21.  Others, including our illustrious leader, think the above position is idiotic, as obviously Wisconsin was not easing up on us.  In his most recent post, Brian mocked those of us in group 1 for just going with our "feelings, man"  instead of presenting any hard data.

But, unless I'm missing something, I've never seen any "hard data" supporting the second theory, either.  Instead of just dismissing the opposing argument as an unsupported "gut feeling," shouldn't people claiming that it's just as hard to score when you're down by 21 as when you're tied actually present some data to that effect?  

I believe in statistics.  I'm not some cantankerous Murray Chass who believes that my gut feeling is more important than newfangled math.  At the same time, I acknowledge my limitations -- I wouldn't know where to begin in gathering the data to demonstrate whether or not defenses "let up" when they're up big.  (If I did, this would be a diary, rather than a board entry).  But I would love to know whom advanced statistics actually support in this area, rather than simply being talked down to by people who claim I don't understand math, but don't present any  numbers in their argument.  Anyone aware of any actual statistical analysis in this area that addresses the question at hand?

Fuzzy Dunlop

July 8th, 2011 at 10:11 AM ^

That's kind of my point, so not sure why you had to begin with an obnoxious "um".  It seems that no one has presented data supporting either position, so I'm not sure why one side is being arrogantly dismissed for lacking statistical support.

But I wanted to make sure I wasn't missing anything, and I'm generally curious about this issue, so I asked if anyone is aware of any analysis that may be pertinent.  Kind of thought that's the type of thing the message board was for.

For those of you making "beating the dead horse" comments -- this isn't solely about the Wisconsin game.  What's past is past.  I'm more curious about the larger issue -- is it, on average, easier to score when you're already down big?  Because of the emotions surrounding the Wisconsin game and last year's regime, it seems that no one has made a fair attempt to tackle the underlying question, which I think is an interesting one.

PeterKlima

July 8th, 2011 at 10:17 AM ^

...how that particular Wisco team handled being "up big."

 

As I mentioned I think if you look at the data/stats/evidence from their other games last year, you will realize it was not "easier" for teams to score on them when Wisco was "up big."  Especially not when Wisco still had their starters in the game.

 

There is also data/stats/evidence that the Wisco OFFENSE did not "let up" when up big in games either.  They were accused of running up the score repeatedly.

 

So, PLEASE look at the Wisco games last year and you will find the statistics you need to determine whether your feelings were right.

 

(While you can aruge the sample size is not big enough, the fact that were are not trying to come up with general numbers, but only how this Wisco team handled itself, means that this data is very significant.)

Fuzzy Dunlop

July 8th, 2011 at 10:25 AM ^

So when some people argue that Michigan's red zone offense was weak, they are told to look at stats indicating that, in general, red zone performance is random.

But when we want to examine the question of whether it's easier to score when down big, we shouldn't look at general numbers, and instead should just look at one team's performance.

Basically, look at a large sample size when that supports your conclusion, but a small sample size when that supports your conclusion.  Doesn't seem particularly intellectually honest to me.

MGoNukeE

July 8th, 2011 at 10:38 AM ^

Most advocates that suggest Michigan's red zone offense was weak say that Michigan only scored on 79% of its red zone possessions, good for 82nd in the country. However, Michigan scored 40 touchdown and 4 field goals in 56 red zone trips, putting Michigan at 24th in the country in points per trip to the red zone and an even higher rank than that if red zone touchdown percentage is used instead. That's hardly weak unless it's only compared to the (few) teams that were higher than Michigan.

Source data: http://web1.ncaa.org/mfb/natlRank.jsp?year=2010&rpt=IA_teamredzone&site=org&div=IA&dest=O

PeterKlima

July 8th, 2011 at 10:53 AM ^

The "red zone issue" acknowledges that UM was bad in the red zone last year, but then discusses whether that performance is an indication of the effectiveness of the spread or other factors.  So, the question is will it be bad in the FUTURE when there is a returning QB and/or FG kicker.  And, the answer appear to be that it should revert to the mean THROUGHOUT THE SEASON because of the randomness when you do not have those "temporary" factors.

 

The "Wisco Scoring issue" is whether Wisco "let up" last year.  It is not about projections.  It is only about ONE GAME and does not accept wether UM did well or badly against Wisco.  In fact, that is the only thing to be determined -- something that happened in the past in one game -- not what are the chances something happens in the future, throughout a season.

 

You need to look at trends to predict outcomes.  You don't need to look at trends to detrermine what happened in the past.

 

Your feeling is that Wisco "let up" on defense ONLY when they were up by 21 to a highly regarded offense.  Go ahead....look at what that Wisco teams performance throughout the season to get your answer.

 

P.S. - The only reason that the "red zone issue" is so much broader is because it is generally accepted it was somewhat bad last year because that is what the season stats show.  But, CERTAIN people can not make the same concession about the offense in general and look at season stats to agree they did REALLY REALLY well.  So, the question is "why can you accept season long red zone stats, but not other season long offensive stats"?  Is it because you have feelings about RR and/or losing to rivals?  "Is it about your feelings man?"

Fuzzy Dunlop

July 8th, 2011 at 11:04 AM ^

So, the question is "why can you accept season long red zone stats, but not other season long offensive stats"?  Is it because you have feelings about RR and/or losing to rivals?  Is it about your feelings man?"

So, you're kind of a dick.  I've done nothing to imply that I'm anti-Rodriguez, that I thought last year's offense sucked, or anything of the kind.  You're the one bringing up dead issues and being obnoxious.

To respond to your points, such as they are:  

The "Wisco Scoring issue" is whether Wisco "let up" last year.  It is not about projections.  It is only about ONE GAME and does not accept wether UM did well or badly against Wisco.  In fact, that is the only thing to be determined -- something that happened in the past in one game -- not what are the cnaces something happens in the future throughout a season.

First, the Wisconsin issue is merely an example -- it was not the only game last year in which our offense came alive after we were down big.  

Second, and in any case, I'm not even saying that our offense last year was bad, or overrated, or anything like that, and I'm not using the Wisconsin game as evidence of our weaknesses. I'm responding to others, who would use the Wisconsin game as evidence of our offensive strength, saying that our numbers in that particular game may have been misleading.

Your feeling is that Wisco "let up" on defense ONLY when they were up by 21 to a highly regarded offense.  Go ahead....look at what that Wisco teams performance throughout the season to get your answer.

Again, you cannot claim that you are conducting a rigorous statistical analysis on this issue using one team's performance as a sample size.  Obviously, there are vast differences in Wisconsin's games against the Michigans and Austin Peay's of the world.  The only reason to focus on Wisconsin's games in particular is because it supports your result.  It's not intellectually honest.

I could just as easily say that, focusing on Michigan's games, it proves that our offense was more likely to score when we were down by 21 than when we were down by 7, and thus it's easier to score when down 21.  The whole point of this post is that I <i>acknowledge</i> that this small sample size doesn't prove anything, and it very well may be the case that our scores in blowout losses are just as significant in assessing our offense.  In other words, I'm asking for data either way to help my understanding of an issue.  While you're just being a dick.

 

 

PeterKlima

July 8th, 2011 at 11:20 AM ^

I just think the Wisco issue is about one game.

 

If you want to have a broader conversation, then why start it by saying the other side claims I have "no data" but where is their data?

 

I mean, you are the one broadening the conversation, so those previous comments about data do not apply to your new conversation and there is no reason to refute them.  Or, to create a thread arguing againt comments that don't have to do with your new issue.

 

I may come off as "dickish" because I feel like you are trying to confuse the issues on purpose, but maybe its just a mistake.

 

Fuzzy Dunlop

July 8th, 2011 at 11:33 AM ^

If you want to have a broader conversation, then why start it by saying the other side claims I have "no data" but where is their data?

Well, in the title of the post I'm asking "who does the data support"?  I acknowledge that my gut feeling may be wrong, and saying that I'd like to see the data either way so that I can understand the issue.

At the same time, I'm expressing my frustration at those who would mock my "gut feeling" by implying that the data refutes it, while not actually presenting any data of their own.  While Brian's recent comments on the Wisconsin game were the impetus for my post, in this regard the Wisconsin debate is really just illustrative of a larger frustration I have with people who act as though all of their positions are supported by hard evidence, yet when you look closer it appears that they are just presenting their own hunches.

I appreciate your toning down the rhetoric in our comments to each other, and I'll do the same.  Calling you a dick (twice!) was over the line.  Sorry.

 

Fuzzy Dunlop

July 8th, 2011 at 10:28 AM ^

Because the Wisconsin issue, raised by Brian in his most recent post, is what brought the subject to mind.  It's called a "jumping off point".

The issue I'm raising plainly has to do with the use of statistics in arguments, and whether one side should accuse others of "lacking data" without presenting data of its own.  I'm obviously not delving into the issue of "was Michigan's offense really any good"?  But if you prefer to be obnoxious rather than consider the actual content of my post, more power to you. 

MGoNukeE

July 8th, 2011 at 10:52 AM ^

Is it really Brian's job to use statistics to prove himself wrong? It should be the responsibility of those that disagree with Brian to show why his statistics don't paint a complete picture of the offense, and show counterarguments backed by statistics that clearly show the offense was lacking.

For example, one such argument is that Michigan turned the ball over too much in Rich Rodriguez's tenure. While this is true, Brian dismisses this by pointing out that we've had a first-year underclassman starter at quarterback every year in Rich Rodriguez's tenure, which has a strong statistical correlation with turning the ball over too much. He expects that Michigan will improve in this category as our quarterbacks becomes upperclassmen and they mature physically and mentally. Until there are stats or other evidence to refute this, those who support Brian aren't going to change their mind about this.

Fuzzy Dunlop

July 8th, 2011 at 11:07 AM ^

Is it really Brian's job to use statistics to prove himself wrong?

No, but it's his job to use statistics to prove himself right.  He believes that points scored in the second half of a blowout loss are as significant in assessing an offense's performance as points scored in the first half of a tight game.  If he dismisses the countervailing theory, he should present some evidence in support of his own.

MGoNukeE

July 8th, 2011 at 12:33 PM ^

The problem with this is that the FEI also dismisses stats accumulated during garbage time. More than likely, Michigan doesn't rank 2nd by their metric if stats obtained after a 24-point halftime deficit are dismissed as garbage. This means that Brian can either default to their method of evaluating gametime as "garbage time" or some other method such as the one emphasized here, where any 21-point deficit is also garbage time.

Is the FEI also evaluating garbage time incorrectly? If you can establish and prove that it is, you will sway many people (perhaps even Brian) on this site.

Fuzzy Dunlop

July 8th, 2011 at 12:53 PM ^

109 posts in, and this is the first useful bit of information in response to my request for actual data.  I wasn't aware that FEI dismisses "garbage time" stats.  Not to give you homework, and I will try to find this myself when I'm done with work and have even more time to waste, but do you know off-hand how FEI defines "garbage time"? 

MGoNukeE

July 8th, 2011 at 2:28 PM ^

No, but I've been trying to find it. So far, Google isn't giving me enough information, but my guess is that it only removes a few drives per game at the end of each half unless it's painfully obvious that a team has stopped trying, like if the NCAA mercy rule kicks in (35-point lead in the 2nd half).

Edit: somewhat unclear explanation from Football Outsiders

"The Fremeau Efficiency Index (FEI) considers each of the nearly 20,000 possessions every season in major collegefootball. All drives are filtered to eliminate first-half clock-kills and end-of-game garbage drives and scores. A scoring rate analysis of the remaining possessions then determines the baseline possession efficiency expectations against which each team is measured. A team is rewarded for playing well against good teams, win or lose, and is punished more severely for playing poorly against bad teams than it is rewarded for playing well against bad teams."

PeterKlima

July 8th, 2011 at 11:00 AM ^

You stuck to the Wisco issues solely in your OP.  When you are confronted with stats from last year that Wisco did not let up in any of its games, then you start to expand the conversation.

 

THAT is pretty evasive and disgenuine.

 

Maybe you can at least concede first that Wisco's stats from last year (with starters in the game) support that Wisco DID NOT let up and UM did pretty well offensively against them (much better than most teams).

 

If you want to "jump off" to a braoder issue, let's make sure you are starting at a reasonable point.  Then, when you jump, think to yourself....

 

DID BRIAN OR ANYONE CLAIM THAT PEOPLE ON THE OTHER SIDE OF THE "BROADER ISSUE" LACKED DATA....OR DID THEY JUST SAY PEOPLE ON THE OTHER SIDE OF THE "WISCO ISSUE" LACKED DATA?

 

(BTW - The "broader issue" is going to include all games last year, etc.  and I think it will be pretty easy when you look at season stats.)

 

 

 

wolverine1987

July 8th, 2011 at 1:03 PM ^

the stats support the fact that Wisconsin did not let up? Honest question. Which stat is it that looks into the souls of the individual players and determines whether or not the were giving as much effort in the 3rd quarter as they were in the first? Or in any game? 

I love how people (not addressing you specifically) feel confident in saying this team or that team didn't let up, or that because the starters were in they were obviously trying their hardest. As if the whole of human history, our own individual natures, examples we can see everywhere in life, from business to sports and in our own houses, do not provide literally millions of examples that yes, in fact, when people feel more comfortable they sometimes relax a bit. And their performance often declines a bit as a result. You can find thousands of quotes just in individual football games where players admit "we let up a bit."  But nope, Wisconsin didn't relax a bit on defense when they were up by 21. Because their starters were in. Ok.

BigBlue02

July 8th, 2011 at 1:54 PM ^

I think he was saying the stats of all the other games Wisconsin was up big in. Unless they let up against us and none of the other teams they were beating badly, I assume that means they didn't give up.

Ziff72

July 8th, 2011 at 10:34 AM ^

I have no data to support this, but 40 years of watching football would tell me emphatically yes it is easier to score when you are way down because the coaches change their aggressiveness due to the fact the game has a clock.  The offense takes more chances which lead to bigger variance.  Typically the defense will change it's aggressiveness to play to the score.

Wisconsin was a good/great team and they beat us last year.   It sucks but in terms of the program who really cares?   Our defense was not ready to compete last year.  If we don't win the BigTen/NC  I don't get caught up in individual games.  I'm pissed during the game and for hours after, but I'm focused on champiosnhips.   This year will be different.

 

jg2112

July 8th, 2011 at 9:44 AM ^

Honestly, the stats support Wisconsin having a much superior, effective offense than Michigan.

In that game, Michigan had 4 drives out of 12 opportunities that I would call "good drives" - a combination of plays, time, yardage and a good result. One in the first half (missed FG the result, but otherwise a good drive) and 3 in the second half.

Wisconsin had 7 out of 12 that I'd classify as the same.

The margin of victory was 20. That's your ballgame.

BiSB

July 8th, 2011 at 9:47 AM ^

Michigan's offense was driving against Wisconsin's defense.  Wisconsin was driving against Michigan's defense.

If Michigan's offense had a chance to play against Michigan's defense, they probably would have had 14 "good drives" out of 12 chances.  Maybe 15.

jg2112

July 8th, 2011 at 10:24 AM ^

I think these are the salient points, and the ones we ought to dive into if we want to dive into anything:

When teams don't have a lot of tape on a Rich Rodriguez offense (the 2009 and 2010 openers, as well as Notre Dame since MICH shut it down in both second halves of the openers), a Rich Rodriguez offense will likely gash you.

If your team is at a significant talent deficit, Rich Rodriguez's offense will undoubtedly gash you. 

If you are a Big Ten coach named Bill Lynch or Tim Brewster, Rich Rodriguez's offense will undoubtedly gash you.

If you are a Big Ten coach with a track record of success, Rich Rodriguez's offense will gash you for the first drive of a game. However, you will be able to outscheme him and contain said offense until half-time because he's a poor in-game mechanic and we don't know what the heck Calvin Magee was doing.

In the second half, once RR can re-group, he'll likely gash you again, but you'll probably have a cushion to protect yourself from said gashing.

BiSB

July 8th, 2011 at 10:29 AM ^

and it also seems to indicate that you're looking at an offense that will pretty much kick ass all of the time, with the exception of the last few drives of the first half against good Big Ten teams.

Aren't we thinking that's a pretty decent track record?

jmblue

July 8th, 2011 at 10:43 AM ^

I would not say that regularly scoring only 7-14 points in the first three quarters of games against quality opposition constitutes kicking ass nearly all of the time.  In seven of our last eight games (the six losses, plus the Purdue win), we followed the same pattern: one, maybe two successful drives at the beginning of the game, followed by a long stretch of being shut down (which usually extended well into the third quarter), before we finally scored again late in the game (if we did score at all, as was not the case against OSU or Miss St).  Aside from the Purdue game, we were pretty much hopelessly behind when that happened. 

An offense that finds itself down 21 points in six of eight games isn't getting it done.  A big deficit can't be blamed on just the defense.  It's the offense's job to keep pace.  The only times it did so in the final eight games were against Illinois and Purdue (and the latter case was due to our defense playing unexpectedly well). 

BigBlue02

July 8th, 2011 at 10:58 AM ^

So you want our offense to score on other defenses at the same pace as other teams scored on our 108th ranked defense? That is ridiculous. Wisconsin had a top 25 defense, Iowa had a top 10 defense, Ohio State had the 2nd best defense in the nation, msu (ntmsu) had a top 40 defense....shit, Illinois had a top 50 defense. Those defenses are all significantly better than ours. It is as though you are suggesting our offense needed to score on 90% of their first half drives or they weren't that good. No offense in the nation did that

BiSB

July 8th, 2011 at 11:10 AM ^

An offense that finds itself down 21 points in six of eight games isn't getting it done. A big deficit can't be blamed on just the defense. It's the offense's job to keep pace

By that logic, these kids share the blame for this loss:

I can say that it's the skinny kid's job to keep up with the other racers, cause if he was REALLY strong, he could make up for the fat kid. But that doesn't mean the blame falls anywhere NEAR equally. At some point, you're just dragging a fat kid around.

Ziff72

July 8th, 2011 at 9:54 AM ^

I have to disagree with you here.  The 2 offenses were pretty similar.  The difference is Wisconsin having a much superior, effective defense than Michigan.

This has become comical.  Everyone says Michigan had the worst defense and special teams anyone had ever conceived of.  We still won 7 games with a 1st time starting QB and 1 meaningful senior and people still want to criticize the offense.

Our defense and expectations the last 3 years has everyone warped.   We went down on Wisconsin and scored 3 td's in a row like it was nothing.   No one did that all year. 

jg2112

July 8th, 2011 at 10:26 AM ^

It's perfectly valid to criticize an offense that does this:

MICH     1st M01  10:53  Punt          M35  08:22  Punt           7-34   2:31

MICH     1st M28  03:15  Kickoff       W13  14:52  Missed FG     12-59   3:23#

MICH     2nd M36  08:07  Kickoff       M40  06:18  Punt           3-4    1:49

MICH     2nd M37  03:50  Kickoff       M40  02:56  Punt           3-3    0:54

MICH     2nd M27  01:51  Kickoff       M27  01:45  Fumble         0-0    0:06

MICH     2nd M15  00:30  Interception  M39  00:00  End of half    2-24   0:30

 

And helps lead the team to a 28-0 halftime deficit.

MI Expat NY

July 8th, 2011 at 12:18 PM ^

You do realize that that fourth "drive" was a fumble on the kickoff return, thus a special teams error, right? 

So, we should blame the offense for producing a chip shot field goal and not much else in 4 drives instead of defense and special teams for us being down 24-0 at half?  This is absurd.  What did you want? Absolute perfection?  Great offenses go four drives without doing much all the time.  If defense or special teams could do anything, maybe we're only down 14-3 by the time the offense sees the field for drives 5-7 that resulted in 21 points.

And just because I don't want to post again in this pointless argument.  We pulled within 10 with 9:45 and 5:31 to go in the third, anyone who thinks that those were meaningless drives is a fool. 

FreddieMercuryHayes

July 8th, 2011 at 9:46 AM ^

This is getting tired. But I will pile on. In the Auburn v. Alabama game last year, Alabama went up 24-0 in the second quarter and utterly shut down Auburn's O. What happened then? Auburn's O woke up (I guess when Bama didn't care anymore) and ended up scoring 28. The difference between that game and ours was that the AU defense also cracked down, forced turnovers and limited Bama to 3 points the rest of the game.
<br>
<br> Was our offense amazing all of last year? No. But it seems like the biggest arguement is that thy couldn't score 75% of the time they had the ball like our opponents could do on us. And that's kind of a ridiculous expectation.

FreddieMercuryHayes

July 8th, 2011 at 11:16 AM ^

AU was down 17-3 against Clemson in the 2nd, down 20-7 vs. SC in the second, was down 6 vs. Arkansas in the fourth and ended up winning by 22, down 21-7 to Georgia in the 2nd (although they did close the gap by halftime) and again won by a bunch. AU made a habit of coming back last year and pouring on the scoring late in games, and it wasn't completely lightning striking. But they won because they had, you know, a defense.

MGoNukeE

July 8th, 2011 at 11:43 AM ^

1) People bring up Auburn-Alabama to show that 2nd half stats matter, and the scoring performed against Wisconsin wasn't some fluke. In fact, had Michigan's defense shown up like Auburn's defense showed up in the 2nd half, Michigan may have done the same thing. Your stat, however, simply says Michigan fell behind by 21 points throughout the entire game in 6 games, but didn't reach a 21-point deficit in 5 of those games until the 2nd half. If you want to cite a pattern, cite the pattern that Michigan was down by 2 scores by halftime in 5 of their last 7 games, or 3 scores in 4 of their last 6.

2) There is no context behind your statistic in that many people will read that and conclude different things about the team. For example, the offense may have needed to keep pace with the opponent's 50-yard TDs by scoring 80-yd TDs. The offense may have needed to score on every possession because the defense was bad enough to let the opposing offense do just that. The offense may have had a first-year starting QB that turned the ball over at inopportune times (like, after gaining 50 yards but before gaining 70 yards for the TD). The offense may have had obtained field goals that were easily makeable but didn't make them since the team lacked a kicker. My suggestion? Use a different statistic altogether, or compile data showing that it truly was the offense's fault for not scoring despite it showing it could gain yards at will against everyone on its schedule.

Ziff72

July 8th, 2011 at 9:47 AM ^

This whole topic is ruined by the fact there are very few who can seperate themselves from their biases against the spread and or Rodriguez.

 

 

 

SirJack

July 8th, 2011 at 10:01 AM ^

It's equally ruined by people who can't separate themselves from their bias in favor of the spread and Rodriguez.

But I think many of us never had a problem with "the spread" per se. I would be perfectly happy with a coach who could bring in the spread and also illustrate he could handle being in charge of the defense and special teams as well.

EGD

July 8th, 2011 at 10:06 AM ^

Well, the dumbest argument being made on this board right now is the one that goes:

(1) our defense was bad last year,

(2) our field goal kicking was also bad last year;

(3) therefore, our offense was also bad last year.

I think Ziff72 is right to say that the people making this argument seem to be predominantly from the "Arrgh spread is femi-ball must run up the gut from power-I formation" camp.  I will say that there are some much more sophisticated critiques of Brian's statistical arguments that are unfortunately tainted by association (involuntary, I realize) with that crowd.