I recently read the book Inhuman Bondage: The Rise and Fall of Slavery in the New World. I highly recommend it. Anyone who wants to compare the situation of NCAA athletes to slavery should really read it or something like it before speaking. http://www.amazon.com/Inhuman-Bondage-Rise-Slavery-World/dp/0195140737
Unverified, Contagious, Outrageous Voracity
Piling on. It's bad when Adult Swim is on your case:
Different kinds of one thing. Braves & Birds makes a good point about oversigning:
The teams that have the greatest incentive to oversign are the middle class or lower class programs that struggle to recruit top players and therefore have to make up with quantity what they cannot acquire in quality. Thus, we would expect that the most successful teams in the conference would not oversign because they don’t have to do so. Therefore, looking at results and recruiting quantities is a fool’s errand because Pennington is not normalizing for program status. In other words, if Florida signs 85 players over a four-year period and Ole Miss signs 105, we wouldn’t expect Ole Miss to have a better record because the extra players will not trump all of the other advantages that Florida has over Ole Miss.
There are two kinds. The first kind (as practiced by Houston Nutt): "maybe if I sign everyone who can play football enough of them will be eligible for me to keep my job." The second: <imperial march> SABAN </imperial march>.
Here's a graph from Brian Fremeau that gives you an indication of just how few kids enroll at Ole Miss relative to the rest of the nation's top 25 recruiters:
Nutt is way down at the bottom with VT, who no one ever talks about; South Carolina, USC(?), and Auburn a bit higher up, then a big band of average followed by places that do not bother with academic issues either because they are morally opposed to skeeze (no one) or don't have to bother (everyone). You'll note LSU and Florida amongst this group. Teams towards the bottom can plausibly argue that their oversigning is less harmful because it consists of signing guys who aren't going to be eligible instead of shoving kids in good standing out the door.
Meanwhile, an SEC partisan is fussin' about Big Ten fans complaining about the competitive advantage provided by oversigning:
Is there some advantage? Sure. SEC teams from 2002 through 2010 averaged 3.42 signees per victory. Big Ten teams average 3.11 signees over the same period. Hardly the night and day difference one would expect.
But while oversigning isn’t the magic bullet Big Ten fans would want you to believe, things like local talent base, tradition and spending serve as tried and true differentiators.
We at MrSEC.com aren’t fans of oversigning. As noted above, we would have no problem if every school went to a hard cap at 25.
But the argument that oversigning is the difference between the SEC and the Big Ten? Well, that doesn’t hold water. And as you can see above, that argument doesn’t even hold water when you make comparisons within the same conference.
Ole Miss throws that entirely out of whack, as do a number of mid-level strivers that are rooting through any large-ish kid in the south to see if any of them can play football.
Meanwhile, I haven't seen many (or any) Big Ten fans say it is the difference between the SEC and the Big Ten. This whole thing is a red herring, anyway: the institutions most publicly against oversigning are Georgia and Florida. Ability to identify skeeze does not stop at the Mason-Dixon line.
It's an obvious advantage that's built on perverse incentives, which is reason enough to get rid of it, differences between conferences be damned. For an SEC fan to rabble rabble about how it's not that big of an advantage on the field misses the point in stereotype-fulfilling fashion.
Terrelle Pryor's lawyer. Is Jackie Chiles:
"It was probably good for Terrelle to meet persons like myself, African-American lawyers, very successful -- quote, unquote," James said. …
"Irrespective of how harsh and idiotic we think some of the NCAA rules are, they are still on the books," James said. "They had slavery for all those years. Those rules are still on the books, and the courts uphold them."
James then ranted about the NCAA and its enforcement process.
"You've got a captured system here in college football. It's mandated, dictated, the student-athletes have no rights. They have no relief."
That is all.
That is not all. Bombs continue to drop on Pryor from all angles. Random NFL GM:
“We spent a lot of time this year going through Cam Newton(notes) and Ryan Mallett’s(notes) personality,” an NFC general manager said. “I haven’t done all my homework on Pryor yet, but my initial impression is that if you line all three of them up and just talked about trust and reliability, Pryor is dead last. Like not-even-out-of-the-starting-gate last.
“And it’s probably only going to get worse.”
Some guy in an otherwise pretty kind Dispatch story:
"People are terrified," Davis said. "They want to really examine the kid as a person, because the stories you hear on the grapevine are not stories that excite you - stories about his leadership, how his teammates respond to him, how he was handled at Ohio State."
And Thayer Evans wrote a story I linked in the sidebar that says an 18 year old male enjoyed having girls send sexy photos to him. I've been on the Terrelle Pryor-emotional-problems bandwagon so long I remember when it was just me and some nuts from Penn State message boards and even I think Evans went a bit too far:
Pryor’s focus consistently led back to one thing: himself.
And while some may foolishly believe Pryor’s statement Tuesday that his decision to forgo his senior year at scandal-ridden Ohio State was out of “the best interests of my teammates,” the truth is that he did it out of selfishness. He did it only to escape being investigated by the NCAA and to try to salvage what’s left of his bleak future.
Well… yeah… but you write for, like, organizations, man.
(Also, Dispatch lol:
The best part about this is the cooler poopers are doing it to themselves.)
Do not read if you are only going to make a tedious argument that shows you don't understand statistics. Bill Connolly, purveyor of Football Outsiders' other college football ranking system and Football Study Hall author, previews Michigan. There are many numbers and a discussion of just how good Michigan's offense was last year (as per usual, advanced stats == fawning) that people who would like to restrict their sample size to four first-half drives against Wisconsin won't like:
If only Michigan had been able to play defense. Despite a slight fade as the season advanced, the Wolverines' offense was incredibly successful in 2010, posting huge point and yardage totals on a series of stellar defenses. Their Adj. Points tell the tale -- against a strong slate of defenses, the Wolverines produced at an incredibly high level for each of the first nine games of the season before a combination of injuries and fatigue (and, possibly, lack of faith in the defense) set in. Michigan still averaged 28.9 Adj. PPG over their final four games, with only two below-average performances against Purdue and Mississippi State.
I'm not sure where Connolly's getting the idea Michigan blitzed on almost every passing down, however. Even if he has numbers for this I kind of doubt them, since I tracked rushers for the Indiana UFR and came up with a ton of 3 and 4 man rushes. If that's charted I wonder if the 3-3-5 threw someone off.
Anyway, Connolly's takeaway is "I hope they don't turn Denard into Brad Smith that one year they tried to make him a pro-style quarterback": since they don't like Nebraska as much as everyone else and their system looks at recruiting rankings that drastically overrate Michigan (attrition) they're hinting the FO Almanac will have Michigan at or near the top of the division.
WTF? I know we're supposed to be taking the high road and all but seriously, if anyone could be expected to jovially bomb Ohio State in the paper it's Mike Hart. Mike Hart:
"I really think Jim Tressel is a great coach," Hart said. "I hate the school, I hate Ohio, can't stand them, but I think he's a great coach. Whatever happened didn't make him a better coach. The players were doing wrong, and (Tressel) broke the rules, which obviously is wrong, but it's not like he was giving them steroids to give them a competitive advantage."
Guh. Multiple other former players say they "genuinely feel" for the other players caught up in this situation who have nothing to do with it, which seems a little much. We're concerned about Ohio State walk-ons and kickers now?
Etc.: Laurinaitis and the Real Girl. Wetzel hears call to argue why OSU's violations aren't as bad as USC's, argues that OSU's violations are worse than USC's. The Wolverine Blog searches for breakout players.
Sorry Brian, the nowhere in the wide world of statistics can you alter reality enough to make Michigan the winner of last year's game. I understand the math and the idea, but it doesn't work for specific games. Michigan got curb stomped, they didn't win by 11(!).
For those who didn't click the link, the adjusted ppg for Football Outsiders has Michigan beating OSU 31.6-20.5.
"ha Michigan really beat OSU." His point is "Michigan's offense was really, really, really good last year, and I hope that doesn't change."
I know, but those stats are just another tool to use in evaluating performance. My point is that they ignore or minimize the effect of things like turnovers, which distorts reality. Some people argue the advanced stats better represent reality and they may for the aggregate, but I don't think it does a better job for specific teams.
Turnovers are a great example. The Football Outsiders guys say that holding onto and stripping the ball are skills, but that falling on the ball is random, so turnovers via fumbles are random. However, if the two acts to get to falling on the ball are skills, then they can be improved and taught, which means that one team can be better at not fumbling and causing fumbles. Teams that are better at those two things will get more chances to pick up turnovers and will have fewer chances to lose turnovers, so given a random probability of picking a fumbleup, teams that are better at those two skills will have better turnover margins.
When Michigan had a -2 turnover margin against OSU last year, that wasn't an accident of chance. Michigan's offense wasn't as good as those stats suggest because they weren't good at holding onto the ball--which I don't think is random luck.
There is no correlation whatsoever between the percentage of fumbles recovered by a team in one year and the percentage they recover in the next year.
Yes, you can teach players to strip the ball, and you can tell them to pursue the ball once it's out ... but that doesn't mean that teaching that skill therefore has a demonstrable impact on the outcomes of games, although recovering a fumble itself does affect the game in a way similar to missing a field goal or dropping a pass.
You may believe otherwise, but that's just, like, your opinion, man.
Michigan's "really, really, really good last year and I hope it doesn't change offense" scored 14 points against OSU, and was shut out after the first quarter. That same "really, really, really good last year and I hope it doesn't change offense" also put up a whoppoing 7 against SMU in the bowl game fiasco (with a healthy Denard and those weeks of extra practices). That same "really, really, really good last year and I hope it doesn't change offense" was shut down for more than half of the game against Penn State's third string defensive backups.
The numbers only show offensive dominance because we put up a metric ton of points against shitty teams. So, I will concede that our offense was "really, really, really good last year" against shitty teams."
I am hoping for change.
aren't talking about points scored when evaluating the offense, because 'points scored" is kind of a crappy measure of offensive capability.
Yes, but while I used points scored in my example, the same is true of all of the other offensive stats. They are grossly overinflated by our dominance of the crappy teams on our schedule.
Michigan moved the ball very, very well on PSU, Wisconsin and Iowa, and moved it fairly well against MSU (they absolutely destroyed the season averages for Wisky and Iowa).
Yes, Michigan put up bigger numbers on Indiana than Wisconsin, but the offense was incredibly effective against pretty much everyone it played.
Were we effective against OSU and MSU (bowl)?
I guess my other problem with the stats is that until they start giving out wins and losses by states other than points, they don't mean much. At the end of the day, our offense showed potential. I am not arguing that. But, I think that it also opened up Denard (who represented approximately 98.325% of the offensive production) to injury. And he was injured in almost every game. It also failed to establish a real non-Denard running game.
Again, there were times when I really liked our offense. I just found that, stats aside, we didn't look as good as some of these stats indicate against better competition.
But we were effective against basically everyone else on the schedule.
And the point of these stats isn't to assign wins and losses after the game has already been played. No one is arguing that Michigan "really" beat Wisconsin or Ohio State or Mississippi State. They're mostly intended to serve as a predictive measure for the future. If you take Michigan's offense over a huge sample, it's going to put up incredibly good numbers because it's very good at getting first downs and moving the ball. And if you pair it with an even remotely competent defense, it's going to win a ton of football games.
I would argue that it's stats like points scored and wins that don't "mean much" from an analytical point of view, because there's both a ton of noise and a sample size issue. Teams can win games despite being outgained by 200 yards (Iowa did it to Michigan this year), but if you had to take one team for eternity, you'd probably want to take the team that gained 200 more yards. In baseball, for example, pitcher Wins and ERA aren't very good measures of pitcher skill (despite the fact that the goal of a pitcher is not to give up runs and the goal of a baseball game is to win) because other peripheral measures are more strongly correlated with long term success. Michigan's offense was just really good at the things within its control that are correlated with long term success (with the possible exception of turnovers, depending on how much control you think a team has over turnover margin).
Michigan's offense was just really good at the things within its control that are correlated with long term success (with the possible exception of turnovers, depending on how much control you think a team has over turnover margin).
According to these advanced metrics, Michigan was really good. However, I disagree that they were really good on the whole. The problem with relying on statistics in a vacuum is that they distort reality. Reality says Michigan wasn't very effective against OSU. They scored seven points, which isn't going to be considered a good offensive performance. The adjusted points based on FO's numbers say Michigan wins on a standard day.
I don't think we're disagreeing that much, I'd agree that points scored isn't necessarily a great metric, but leaning on the stats to prove Michigan was really good doesn't make much sense either, given reality.
If you take both stats and points scored into account, hopefully we can all agree that, while the offense was not as effective as we thought from the first games (in terms of scoring against good teams), it had incredible potential and was dangerous. I think saying that the offense did not work, therefore we needed change is just a major oversimplification. If there was one thing that was working last season was the offense (yes, I know, except for OSU and MSU, ntMSU).
To say that the offense wasn't working and stop right there, ignores that there were several reasons why it didn't perform to it's potential:
- Denard as a first year starter. In spite of that, Big Ten Offensive POY.
- Michigan offense needed a real alternate running threat. Once teams realized Denard wad the only running threat, they knew who to focus on.
- Michigan defense and special teams were disastrous. As an offense, when you know that you know that the only real chance at winning is to score a TD in almost every possession, it creates an incredible amount of pressure that only an offense filled with upperclassmen could handle.
Yes, the Michigan offense did not do enough to win key games, but in my opinion, the reasons above (not excuses), especially the defense and special teams, made "enough" a goal nearly impossible to reach. At the end of the season, I wanted change too. Change everything, defense, special teams. But at the risk of sounding like a grumpy old man, "get off my offense!" (lawn).
Compare how Michigan's Offense performed against Ohio State, Michigan State, Iowa, etc and then compare how other offenses performed against those defenses. That is what the FO method teases out. M's offense was good and hadn't even reached its full potential yet.
Your points about Vicotries are valid but those are the result of the effort of the entire team not just the offense. Michigan had most of an offense, half (at best) of a special teams unit, and zero defense in 2010. Sum that up and it doesn't add up to enough against quality opponents. It's not that hard to understand if you actually care to do so.
I'm not sure if you're replying to me or not.
I think they're attempt is nice, but leaving out really significant events like turnovers and calling them random distorts the stats and makes a team like Michigan look a lot better than they actually were. Fumbles aren't random events.
I explain myself in my earlier post.
I've speculated on this subject very extensively in the past (link)and--for the most part--I agree with you, but while I agree that forced fumbles aren't completely accidental, I believe that fumble *recoveries* are definitely random. So, a skilled defense will end up with a lot of fumbles recovered because they generate a lot of opportunities then get their fair share of those. I haven't found a solid data source to test that hypothesis yet, but that's what I think.
Offensively, I'd venture that maturity has a lot to do with it (i.e.: freshman fumble more than seniors). Again, can't prove it but I'd like to test that theory with solid data.
Stats get misused all the time to make shitty arguments, you're right about that. However, its hard to argue that Michigan wasn't more successful (on a wholistic basis) than most offenses were. The arguments against Michigan's 2010 offense amount to a complicated version of "they weren't as good against good defenses" which, um, duh...that statement is true about all offenses.
"Is there some advantage? Sure. SEC teams from 2002 through 2010 averaged 3.42 signees per victory. Big Ten teams average 3.11 signees over the same period. Hardly the night and day difference one would expect."
This is like the weirdest made up stat ever to evaluate this situation. How did he decide this stat taken from an entire conference (not all of which oversigns) would be the best way to evaluate the data? I have no idea.
Also, despite the fact that he jams the number into the smallest units he can (it's only a 0.3 difference! Who could be worried about 0.3?) it is still a 10% difference.
So yeah, even though I don't understand why he picked that ratio, if we accept it measures what he wants it to for arguments sake, who is dismissing a 10% advantage?
Seems pretty huge to me.
28.9 Adj points per game. Scoring below average only against Purdue and Miss St. What did Michigan score against OSU?
makes the best argument I've read yet why OSU deserves the full hammer from the NCAA. He also brings in Delaney, stating that when Delaney signed off on the "punishment" at the time of the bowl game he should have known that the "investigation" seemed like a white-wash. Very good stuff.
Unfortuately for all of us last year advanced statistics =/ wins on the football field. The offense didn't do it when it mattered.
between an offense that was "clutch" (which isn't a thing) but had pedestrian peripheral numbers, or an offense that "didn't do it when it mattered" but had outstanding peripherals, I'd take the one that underperformed every day of the week and twice on Sundays.
No I gave you a choice between an offense that wins football games and one doesn't, shouldn't be that hard of a choice.
doesn't actually tell me anything at all about whether I'd want it to be my offense in the future. Les Miles's offense won more football games last year than Michigan's did. Would you have traded LSU's offense for Michigan's?
We're discussing Michigan's offense here and justifiying their performance using convoluted metrics. The offense isn't the only thing that adds up to wins so this comparison to LSU is irrelevant. You'd prefer Oregon's offense to ours probably right? Probably because they won more games.
The point is we had an inconsistent team in all phases last year. I don't care what metric you use the final scores speak for themselves. Plenty of Michigan teams in the recent past should point to a less than explosive offense getting the job done just fine. I'm not going to be picky about the scheme if we start winning games.
because they were more effective at moving the ball down the field (I think they're the team that was ahead of Michigan in FEI right?). In terms of just evaluating how good an offense was, I don't care all that much how many points they scored or how many games they won (obviously I care about that stuff from a fan perspective, but not from an analytical one), because I don't believe those are very good measures of the quality of an offense. There's too much noise in them.
But statistics can say whatever we want them to, I don't understand what the point of going that deep is? College football is so fluid and the quality of opponents so varied I don't see how there is any less noise in these numbers than others.
I guess my main problem is that the "if only Michigan had a defense" line pisses me off because I definitely saw the offense falter plenty last year. They were absolutely the reason we won most of our games but they were also culpable in our losses. Using stats to minimize the way we lost those games is wrong. At the end of the day the offense had some wonderful stats but the team didn't win the games.
is to isolate the factors that play the biggest role in winning and losing football games. Brian's point is that Michigan was really good at those factors on the offensive side of the ball.
I guess I just can't buy that compared with the results of the games and what I saw on the field. Maybe being shutout in 6 of the last 8 quarters of the season has tinted my opinion of the previous games. But against Wisconsin and MSU and Iowa (until we abandoned our offensive plan and just chucked the ball) our offense was absolutely held in check until those teams had leads over 20 points. The defense was certainly a problem in those large deficits but the offense did not do much to keep those games close.
I am amused that the football outsiders guy attributes the dive in offensive performance over the last four games to fatigue, and ignores the rather obvious uptic in opponent quality (OSU, Wisconsin, Mississippi State)
Did you write "The Jungian Thing" over at the WLA site? If so, well done. Should be mandatory reading for those looking at the Tressel situation.
Waving away a quarter of your sample, against the best competition, is a giant red flag.
I'm not sure what it means but it definitely leaves room for skepticism.
Talking about signees per victory doesn't make sense. In fact, if the correlation between number signees and victories is strong, you would expect all teams to have the same number of signees per victory.
Secondly, the record of an entire conference gets dragged toward .500 by the mathematics of the schedule. About 65% of an average Big Ten or SEC team's games (depending on bowl eligibility) are played within the conference. So the conference's record in those games is, by definition, .500. Any diffference in victories comes from bowl games and the early-season nonconference cupcake schedule. That is a miniscule difference. Last year, the Big Ten was 82-58, for a winning percentage of .586. The SEC was 95-61, for a winning percentage of 0.609. Had both conferences played the 156 total games that the SEC did, the Big Ten's percentage prorates to 91.4 wins compared to the SEC's 95, for a difference of 3.7%. Conference-wide, it is very difficult to have a significant difference in record.
Also, all teams have 85 scholarships available, so every team should be giving out around 20 scholarships per year unless there is significant attrition. That really washes out the correlation between signees and victories, because Eastern Michigan gives out the same total number of scholarships per year that Auburn does.
The only thing that the signees per victory stat tells me is that, in the last ten years, the SEC has signed about 10% more players than the Big Ten has, which means that their attrition is significantly higher. We can all agree that that is not a good thing.
I have to guess that the 3-3-5 is creating semantic problems with the word blitz. It happens all the time with the Steelers and their 3-4. In my book, a blitz is sending more than 4 guys but people see a LB rushing and call it a blitz. But when you only have 3 lineman the 4th guy has got to be a LB or DB--that doesn't make it a blitz, though.
Although, I would have thought Connolly would have been wise to this.
Although, I would have thought Connolly would have been wise to this.
Or he has an unknown (to us) reason for calling it a blitz even though there's only 4 rushers. I'd look into this, but I'm lazy.
"Guh. Multiple other former players say they 'genuinely feel' for the other players caught up in this situation who have nothing to do with it, which seems a little much. We're concerned about Ohio State walk-ons and kickers now?"
I have an intense dislike for OSU, the vast majority of their fanbase, and their athletic administration (consistent with many others around these parts). But like Hart and other ex-players commenting on the situation, I agree that it's an unfair situation for the kids on OSU that did not partake in the free ink / free car perquisites. Notwithstanding their terrible taste in universities, those kids are getting screwed by their teammates, and IMO deserve some compassion.
Has Mike Hart stated an opinion on practicegate? Couldn't find anything on the interwebs! Sorry to slant the subject - inquiry minds would like to know, ya know!
Great post Brian...And yes, Wexner is the long reigning unofficial mayor of Columbus. I'll say this, that Hilton Hotel he owns has some of the best pancakes! Nomnomnom!
I have been one of the loudest critics on this board of Brian's vocal displeasure every time a former player opens their mouth. And I have ALWAYS defended Mike Hart.
HOWEVA, I must join in the "guh, Mike Hart sentiment."
When Mike says, "Whatever happened didn't make him a better coach," he is just flat out wrong. PLaying ineligable players gives him an advantage. Whereas RichRod (who I am also a vocal critic of had the decency to suspent Hagerup in the OSU game - arguably the game in which he was fighting for his job - Tress played the Tat 5. That gives him an advantage. Not to mention to possible recruiting violations that may have contributed to his immense talent haul every year. That gives him an an advantage and creates the PERCEPTION that he is a better coach because he wins.
Mike Hart not talking trash about OSU might mean that he has learned a karmic lesson from the "Little Brother" quote. The 'politically correctness' of his quote also could indicate his interest in coaching in the future.
Every time Mike Hart opens his mouth I die inside a little. I wish he would keep his mouth shut. Everyone has opinions but not everyone needs to open their yappers and express them, especially when the topic does not concern them. I'll never understand why "public" figures feel the need to weigh in on things like this. NO good can come of it!
highly touted secondary? Why do I not recall this development?
It has come to my attention from reading this thread that some of you did not observe Brian's paragraph heading, to wit:
Do not read if you are only going to make a tedious argument that shows you don't understand statistics.
Please make note of it in all future correspondence.
The points are these
1. Michigan had a lethal offense last year, by all advanced metrics that account for confounding factors that make it difficult to figure out whether an offense was truly good or not.
- Stipulating this does not constitute criticism of Brady Hoke.
- Stipulating this does not constitute a comprehensive argument against the firing of Rich Rodriguez, largely because
2. Michigan had an atrocious defense last year. Their defense was as bad as their offense was good.
Now, what many people who do not understand statistics will say is "but, but, but, 7 points against OSU!!!" What they are not understanding is that 1 and 2 above are related. That is, it is difficult to score more than 7 points against OSU when your defense does not give your offense the ball back in a timely fashion. Having a good offense does not mean that you score every time; hence the more opportunities a good offense has to score, the more points it will score.
The reason FEI is a valuable statistical tool is that it, among other things, eliminates the confounding factor of #2 in assessing #1.
So, to recap:
- M had a really good offense last year.
- M had a really bad defense last year.
- With a better defense, the offense would have put up more points.
- But the defense was really bad, so the really good offense didn't do this.
- The combination of really good offense and really bad defense led to a very nearly .500 team, which is sort of what you'd expect, no?