Thank you. I, too, was frustrated with what I perceived as cherry-picking data, and I'm glad we can find some common ground.
I think in broad terms, we have some agreement. A win expectancy of 20% isn't crazy, and if it matters (at least in terms of pregame perceptions), Vegas odds put Michigan at around a 22% chance of beating OSU this year.
I also agree with the general premise that Michigan will struggle to beat OSU with the differences in recruiting, and I do think Harbaugh is a good coach, but I also feel that we should expect better than an 0-5 record. All of those things can be true.
Finally, just as Stanford and USC may not have had completely equivalent coaching during Harbaugh's run, I'm not sure that the entire Michigan staff has been completely equivalent to the entire Ohio State staff over the past 5 years. Just as football talent isn't all about the star ratings, coaching isn't all about the head coach. In the end, though, Harbaugh has to own the whole program and the overall results.
Look, my objection has always been the claim that equivalent coaching with a talent differential translates to a 10% win expectancy.
I may not be explaining my point well, but the calculation of the probabilities is straight-forward. For Stanford to go 3-1 against USC given a 10% win expectancy, is to claim that what Harbaugh did at Stanford was a 1 in 270+ event, or something that would happen only once in 1100+ years.
The formula is (n choose k) * p^k * (1-p)^(n-k) where n is the total number of games, k is the number of wins, and p is the win expectancy. 0-5 at 10% win expectancy is (5 choose 0) * .1^0 * .9^5 = 59%. However, 3-1 at that same win expectancy is (4 choose 3) * .1^3 * .9^1 = 0.36%.
To me, it's much more reasonable to say that equivalent coaching and talent differential has something like a 20-35% win probability. At the low end of 20%, Stanford going 3-1 against USC would be about a 1 in 40 event (once every 160 years), and Michigan going 0-5 against OSU would be about a 1 in 3 event (once every 15 years). At the high end of 35%, both situations would be about a 1 in 9 event (once in 36 years for Stanford, once in 45 years for Michigan).
Why would I use a range of win exoectancies? Because I expect significant variance based on many factors, not just talent alone. What you call luck, or some might call bad officiating, or a lack of execution, or a lack of discipline, or weather, or any other factor that would influence winning a football game - all of that would translate into significant variance in win expectancy - even assuming equivalent coaching. However, if you choose to collapse every factor into a single number, 10% does not appear to be a reasonable assumption in that it would imply Harbaugh had god-like powers at Stanford. I don't think that's reasonable.
What I do think would be a reasonable statistical claim is that a more talented team is more likely to win in any given year, but a less talented team is still unlikely to go 0-5 against it's more talented opponent, assuming equivalent coaching.
Let's take your hypothesis - "football is dominated by gross motor skills" and "height, weight, speed...traits of athleticism" are the primary determining factors in a football player's "talent."
What would be the null hypothesis? Football is complex, and factors other than "gross motor skills" and "height, weight, speed...traits of athleticism" contribute to a football player's "talent."
Can we rule out the null hypothesis? No. There are too many examples that support the null hypothesis - higher rated recruits that "bust," lower rated recruits that break out, track stars that are lousy receivers, and slow receivers that are stars (do you know Jerry Rice's 40 time?). Clearly, there's more than just "gross motor skills" and "height, weight, speed" that makes for successful football players.
Scheme matters. Coaching matters. Situational awareness matters. Complementary football matters. And yes, other skills besides "gross motor skills" and "height, weight, speed" matter, even for offensive linemen - skills like hand technique, foot placement, stance, balance, communication.
But even if we accept your hypothesis, despite the evidence to the contrary, and we accept that the more talented team will win a lot more, that still doesn't justify 0-5. Literally the only way that an 0-5 record should be expected is if Ohio State is more than 8 times better than Michigan (> 8:1 odds of beating Michigan in each game). Not even Vegas believes that.
We shouldn't expect that and we shouldn't settle for that.
You created a post claiming to provide a "Data-Driven" approach, and then you use horrible statistical assumptions and methods in your approach. I've shown you how your calculations are faulty, and you continue to dismiss data that doesn't agree with your predefined conclusion as "blips" or "lucky" or outliers.
That's horrible "Data-Driven" methodology, and I won't apologize for calling it out on a post with "Data-Driven" in the title.
The reality is that 0-5 is the statistical outlier, not the expected trend. We should expect better.
I'd also point out that Stanford and USC have been 50-50 since Harbaugh left, despite USC consistently having higher rated recruiting classes. There's your trend.
I agree with your latest point - that coaching does matter. In fact, I would argue that coaching and program building matter at least as much as raw talent, if not more. Furthermore, I would argue that in no reasonable comparison of Power 5 conference football teams, would talent dictate wins in the way you suggest.
I like to use the example of Stanford vs USC when Harbaugh was at Stanford, because it negates a lot of the preconceptions that criticism of an 0-5 record vs OSU is automatically an unfair criticism of Harbaugh.
Stanford under Harbaugh went 3-1 against USC under Pete Carroll. I think it's reasonable to suggest that Pete Carroll is not a significantly worse coach than Jim Harbaugh. Yet Stanford, with more stringent academic requirements, significantly lower rated recruits than USC, and a backup QB in the 2007 game, won 3 out of 4. In that situation, Harbaugh and his staff built a better program and coached the players more successfully than Carroll and his staff.
To now suggest that an 0-5 record against OSU is a reasonable expectation for Michigan flies in the face of actual evidence, including evidence of Harbaugh's own excellent work in the past. Harbaugh wasn't hired to recruit several spots lower than OSU and lose the Game every year. Harbaugh was hired because he had a demonstrated track record of winning against other top coaches and against more talented teams - even with stringent academic requirements and a clean program.
We have to accept that while Harbaugh has proven to be a good coach in the past and even in the present, he and his staff need to do better now and in the future. The lack of effective DT players is partially Harbaugh's responsibility, the lack of discipline and execution is partially Harbaugh's responsibility, and the 0-5 record is partially Harbaugh's responsibility. He has to own it, and we have to expect him to own it. Talking about a talent discrepancy as if it defines football destiny is simply not valid, statistically or otherwise.
I've said it before, football is notorious for small sample sizes. You have 5 data points and are arguing for a statistical correlation. Yet you could take a different set of 5 data points (say MSU vs OSU 2011-2015 or Stanford vs USC 2007-2011) and find an inverse correlation.
I have two problems with the general argument that more talented recruits somehow justifies the 0-5 record against OSU. First, 0-5 is statistically very different than any other record. If Michigan should lose 70% of the time because of talent, then an 0-5 record still only has a 16.8% probability. Winning at least one game has a 36% probability. Put another way - even if OSU's talent means it should win 70% of the time, any record other than 0-5 has a probability of 83.2%. Going 0-5 is the outlier - it should not happen. It should not be expected.
Second, the talent = wins argument devalues program building and coaching, yet we see the impact of program building and coaching, and the limited impact of talent, in every sport. There's plenty of contrary evidence of less talented teams beating more talented teams, and if it really all boils down to talent, then what's the point of hiring good coaches? Of having good schemes? Of complementary football?
We have to value program building and coaching, and we have to believe (and expect) that talent alone will not dictate what happens. We should expect better.
Your claim wasn't that we needed to be within 2-3 recruiting spots of Ohio State to be 50-50, your claim was that being more than 2-3 spots lower than Ohio State would result in 10-90.
If the recruiting difference between Michigan vs Ohio State truly results in a win probability of 10%, then the probability of an 0-5 record is about 59%. OK - Michigan is doing what's expected. However, that also implies, by your same standard, that the probability of MSU beating Michigan 7 out of 8 times is something like 0.0000009%.
By contrast, if recruiting rankings are not so powerful in determining the outcome of games, but are only an influence (say, 30% win probability instead of only 10%), the probability of going 0-5 against OSU is only 16.8% (and the probability of MSU winning 7 of 8 is still only 0.015%).
Football is notorious for having small sample sizes, but that works both ways in your argument. You make an implicit statistical argument - we shouldn't expect anything better than 0-5 because the probability of beating a team with more than 2-3 spots higher recruiting rankings is only 10%. Then you reject contrary evidence (Auburn, MSU, Stanford) as statistical outliers ("lucky events"). But if those are outliers, then Michigan's 0-5 performance is also an outlier - it shouldn't be happening. It shouldn't be expected.
Michigan will – in all likelihood (e.g., 90%+ probability) lose every year to Ohio State unless we recruit better (defined as being within 2 – 3 spots of OSU through the 247 Sports composite rankings)
I appreciate the analysis that went into this, but my question is whether Auburn has been consistently within 2-3 spots of Alabama in the 247 Sports composite rankings since 2013, when they beat Alabama 3 out of 7 times. Or whether Stanford was consistently within 2-3 spots of USC in those recruiting rankings from 2007-2010, when they (under Harbaugh) beat USC 3 out of 4 times. Or whether Michigan State was consistently within 2-3 spots of Michigan in those recruiting rankings from 2008-2015, when they won 7 out of 8.
If you were to conduct an analysis of recruiting rankings over time, is it really true that teams 2-3 spots higher in those rankings will win 90% of the time?
Were 2018 and 2019 worse defensive performances than 2015, 2014, and 2013?...It took Ohio State only 7 possessions to reach 42 points in 2015...Our defense on Saturday didn't fair a lot better than that 2015 group did, but after 9 possessions, OSU had 42 points.
Here's the problem with your analysis. Time of possession for Ohio State:
* 2013: 33:21
* 2014: 25:41
* 2015: 29:54
* 2016: 28:47
* 2017: 28:43
* 2018: 24:36
* 2019: 23:35
Michigan's defense gave up more points in less time over the past 2 years. So saying that OSU had more possessions is not a good thing for the defense - it means that OSU was able to score more quickly, more easily, with less time. You've proven that OSU was able to cut through the defense like a "hotter" knife through butter - that's not a compliment to the strength of the butter.
1. The offensive game has evolved rapidly over the past five years, and even the best defenses are giving up more points than ever before.
That would be comforting if Michigan's defensive performance was comparable to "the best defenses" against Ohio State. The problem is that Michigan's defense gave up:
* More points than Florida Atlantic, Cincinnati, Indiana, Nebraska, Michigan State, Northwestern, Wisconsin, and Penn State, and gave up the same number of points as Rutgers.
* More yards than Florida Atlantic, Cincinnati, Indiana, Nebraska, Michigan State, Northwestern, Wisconsin, and Penn State.
* More yards per play than Florida Atlantic, Cincinnati, Indiana, Michigan State, Wisconsin, and Penn State, and the same number of yards per play as Maryland.
And that was with Ohio State possessing the ball for only 23:35, compared to their season average of 31:38
3. Brown adapts. My estimate from watching games (far different from what announcers say) is that we played about 50:50 zone vs. man coverage. We shut down crossing routes. He's keeping our defense modern and adaptable.
Except Parris Campbell is not on the 2019 Ohio State team. Ohio State's 2019 offense goes through JK Dobbins. Scheming to stop a player that's not on the team anymore, and failing to scheme to stop the key player on the current team, is not a positive adaptation.
If Harbaugh leaves, I'd honestly like him to take the head coaching position.
No. How could anyone justify promoting Brown to head coach after the last 2 defensive performances against Ohio State?
I think it will take a special QB. That's a great equalizer. Where is Stanford without Andrew Luck? Where is Michigan State without Kirk Cousins or Connor Cook? Purdue without Drew Brees? Even when Purdue beat OSU last year, we've seen that David Blough is at least a backup-level NFL QB. Unfortunately, I don't think Shea Patterson was special enough.
While I agree that NIL rights will change things, I disagree that the playoff system didn't change anything. The playoff system not only provided more clarity about the national championship, it also focused more attention on the national championship as THE goal. Prior to that, getting to the Rose Bowl (or equivalent New Year's Day bowl) was considered the highest reasonable goal for any program other than Alabama, because the national championship was left to the unpredictable vagaries of "voters" and "media bias" and "scheduling."
Before the playoff system, recruiting between OSU and Michigan could be somewhat balanced, because neither program really had any reasonable claim to playing for a national championship, and both programs could provide equivalent exposure on New Year's Day. If anything, Michigan under Harbaugh might have been able to provide better exposure.
OSU winning the first playoff while Michigan went 5-7 under Hoke really established a clear distinction between the programs at a crucial time.
Big10 championship down to the wire. One goes to the playoff, and Pac12 champ is left out. Said Pac12 champ then destroys Big10 runner-up in the Rose Bowl, and Big10 champ destroyed in the playoff. No way the committee puts 2 Big10 teams in *at the cost of the Pac12 champ* this year.
Questionable Big10 champ with close win over Big10 runner-up gets in over Pac12 champ, who promptly unpants said runner-up in Rose Bowl? I don't think the committee freezes out Pac12 champ 2 yrs in a row in favor of a 2nd Big10 team.
The reason Brady was suspended 4 games is because of the Patriots' opponent in their 5th game. Imagine the TV ratings...
And Brady wasn't asked to hand over his phone. He was asked to choose what texts were relevant and print them up for the investigators, and they would trust his judgment about what was relevant. He declined to do even that much.
There's a perception, accurate or not, that star athletes get special treatment by law enforcement. That perception then leads to a presumption - when law enforcement does take action against a star athlete, it is probably an under-reaction rather than an over-reaction. Hit-and-run turns into a traffic ticket, body slamming a security guard becomes a "Youthful Trainee" sentence, whipping a child until he bleeds becomes a no contest plea to a misdemeanor, etc.
Frank Clark's arrest occurs in this context. We can all advocate for innocent until proven guilty, waiting for the legal process to play out, making sure all the facts are in before jumping to conclusions, etc. However, I think we also have to recognize that the law isn't always blind when it comes to star athletes, and we can't always rely on the outcome of a legal process to completely define how we should view, or judge, a particular situation.
It's fair to hold our legal system to a standard that the accused are innocent until proven guilty. It's fair to hold reporters to a standard of framing their reports with the word, "alleged." But I also think it's fair for the public, including people who post on fan blogs, to view events like this within the context of how star athletes are treated by law enforcement in our society.
I hope we're willing to recognize that "running up the score" is inappropriate even when it's done by someone (we want) on our side. However, just as some want to down play the "spear" because it was done by our players, I also suspect some are complaining about the last TD because we weren't the ones scoring it.
MSU went up 21-3 on Gardner's pick 6 - that wasn't the defense's fault. Except for the drive at the end of the first half, I thought the defense did a decent job with how little help it was getting from the offense - 2 net yards in the first half, only 10 minutes ToP, etc.
for most of the first half. Giving up a TD on the last drive of the 1st half was a killer. But without some life from the offense, the defense couldn't keep it up.
Michigan 1st half drives:
4 plays, 17 yards, punt, 2:01, 0-7
1 play, -9 yards, fumble, 0:07, 0-7
9 plays, 25 yards, punt, 5:46, 0-7
3 plays, -3l yards, punt, 1:43, 0-7
3 plays, 0 yards, FG, 0:48, 3-7
At that point, MSU had a 2:1 advantage in ToP. And Michigan had gained all of 2 net yards. Any defense would be hard pressed to keep up the fight in that situation.
There's more to Michigan's reputation than athletics. How Shane Morris was handled does matter. Anyway, I'll post in Ace's thread - I had seen this one first (hence the edit).
Yes, the regents probably could do with a little more sunshine. But realistically, if the other regents believed that anything they said could end up posted on MGoBlog, they would stop talking to Brian except in the formal meetings. The closed doors would just move somewhere else, and exclude Brian.
Also, Brian would probably be forced to start editing himself on MGoBlog to avoid creating legal liabilities for the University. As a regent, he would have a duty to the best interests of the University, not to us, and would have to honor things like FERPA and HIPAA, legal counsel's opinions, etc. Then we, the readers of MGoBlog, would stop trusting Brian as much because we would sense he's not telling us everything.
It's very tricky to try and be both an outside advocate and an inside policy maker. Usually, it doesn't work out very well for the person in the middle.
Maybe going for 2 right away is seen as being aggressive and trusting your offense to get the job done, and trusting your defense to get an extra stop if necessary? Feelingsball can always go in multiple directions.
Win or lose, I'll be looking for sustained discipline. I think that playing with emotion, but keeping those emotions under control, will be the most challenging part of the game this Saturday. Given all the distractions and emotions of the week, I could see emotions running high but losing discpline (penalties, miscues, etc.). I could also see emotions running low, and a flat performance or folding in the face of adversity.
Even if they lose, if they appear to keep their emotional edge but stay in control, that will be a good sign for the direction of the team.
Even if they win, if they appear to be undisciplined or flat, that will be a bad sign for the direction of the team.
If people were calling for Brandon to be fired by claiming he slept with a student assistant, or claiming he embezzled funds from the University - that would be a witch hunt. Calling for him to be fired for factual aspects of his job performance is not a witch hunt. You can debate whether his job performance merits being fired or not, but you can't dismiss those calling for his ouster by claiming they are conducting a "witch hunt."
Let's imagine 2 different drivers are pulled over for speeding by the same police officer on the same road going the same speed. One has a clean record, immediately takes responsibility and commits to ensuring it doesn't happen again. The other has a record of multiple traffic violations, reacts defensively to the officer, and insists that it was the passenger's responsibility to keep an eye on the speedometer.
While both drivers might receive speeding tickets, we can understand if the officer chooses to issue a warning to the first driver, or if the court applies a more lenient sentence to the first driver.
Context does matter, and it's unfair to the protesters to insist that they should always treat each event in isolation and ignore the context.
An alternative might be that Morris didn't have a concussion, but they're concerned about admitting that they didn't follow the concussion protocol at the appropriate time.
If they evaluated Morris after the game and found that he had no concussion, it would still cause some potential problems (including liability) if it was revealed that the evaluation didn't happen until after the game.
Still, there might have been a way to finesse the language to obfuscate the timing. So if he was evaluated after the game and found not to have a concussion, there probably would have been a way to get the good information out (no concussion) without being too specific on the bad information (evaluated after the game).
So, that's just long way of saying your explanation is most likely accurate.
Just answering your first question - they could announce a review of player safety protocols by a credible 3rd party or by a credible faculty group. Universities have structures and protocols for all sorts of situations that have ethical questions, ie research involving human subjects, etc. It doesn't have to be an extraordinary thing for a University to study itself.
"...we ran out a coach who had improved each year but not fast enough for us. He has, of course, become the head coach at another program, improved that team each year..."
A Power5 conference school is struggling and hires a new coach:
Head coach at another Power5 conference
Turned around that program from 12 wins in 4 years to 5 consecutive bowl games
Led that program to its first BCS bowl game in school history
Won Coach of the Year in his conference the year he's hired away to new position
After only 2 years, he's fired. The AD gets a lot of heat for firing a proven coach so quickly.
"This decision...was a result of not seeing the progress in some of the critical areas we would have hoped for after 2 years...I think the most critical assessment I had to make was, does investing another year present the likely possibility of substantial improvement? If I couldn't answer that affirmatively, which I ultimately didn't, I felt it was better to make the change now than wait more time."
The AD? Bob Bowlsby. The fired coach? Walt Harris. The school? Stanford. The replacement? Jim Harbaugh.
Recent Comments
Thank you. I, too, was frustrated with what I perceived as cherry-picking data, and I'm glad we can find some common ground.
I think in broad terms, we have some agreement. A win expectancy of 20% isn't crazy, and if it matters (at least in terms of pregame perceptions), Vegas odds put Michigan at around a 22% chance of beating OSU this year.
I also agree with the general premise that Michigan will struggle to beat OSU with the differences in recruiting, and I do think Harbaugh is a good coach, but I also feel that we should expect better than an 0-5 record. All of those things can be true.
Finally, just as Stanford and USC may not have had completely equivalent coaching during Harbaugh's run, I'm not sure that the entire Michigan staff has been completely equivalent to the entire Ohio State staff over the past 5 years. Just as football talent isn't all about the star ratings, coaching isn't all about the head coach. In the end, though, Harbaugh has to own the whole program and the overall results.
Edit: I'd like to de-escalate the tone.
Look, my objection has always been the claim that equivalent coaching with a talent differential translates to a 10% win expectancy.
I may not be explaining my point well, but the calculation of the probabilities is straight-forward. For Stanford to go 3-1 against USC given a 10% win expectancy, is to claim that what Harbaugh did at Stanford was a 1 in 270+ event, or something that would happen only once in 1100+ years.
The formula is (n choose k) * p^k * (1-p)^(n-k) where n is the total number of games, k is the number of wins, and p is the win expectancy. 0-5 at 10% win expectancy is (5 choose 0) * .1^0 * .9^5 = 59%. However, 3-1 at that same win expectancy is (4 choose 3) * .1^3 * .9^1 = 0.36%.
To me, it's much more reasonable to say that equivalent coaching and talent differential has something like a 20-35% win probability. At the low end of 20%, Stanford going 3-1 against USC would be about a 1 in 40 event (once every 160 years), and Michigan going 0-5 against OSU would be about a 1 in 3 event (once every 15 years). At the high end of 35%, both situations would be about a 1 in 9 event (once in 36 years for Stanford, once in 45 years for Michigan).
Why would I use a range of win exoectancies? Because I expect significant variance based on many factors, not just talent alone. What you call luck, or some might call bad officiating, or a lack of execution, or a lack of discipline, or weather, or any other factor that would influence winning a football game - all of that would translate into significant variance in win expectancy - even assuming equivalent coaching. However, if you choose to collapse every factor into a single number, 10% does not appear to be a reasonable assumption in that it would imply Harbaugh had god-like powers at Stanford. I don't think that's reasonable.
What I do think would be a reasonable statistical claim is that a more talented team is more likely to win in any given year, but a less talented team is still unlikely to go 0-5 against it's more talented opponent, assuming equivalent coaching.
Can we agree on that?
Let's take your hypothesis - "football is dominated by gross motor skills" and "height, weight, speed...traits of athleticism" are the primary determining factors in a football player's "talent."
What would be the null hypothesis? Football is complex, and factors other than "gross motor skills" and "height, weight, speed...traits of athleticism" contribute to a football player's "talent."
Can we rule out the null hypothesis? No. There are too many examples that support the null hypothesis - higher rated recruits that "bust," lower rated recruits that break out, track stars that are lousy receivers, and slow receivers that are stars (do you know Jerry Rice's 40 time?). Clearly, there's more than just "gross motor skills" and "height, weight, speed" that makes for successful football players.
Scheme matters. Coaching matters. Situational awareness matters. Complementary football matters. And yes, other skills besides "gross motor skills" and "height, weight, speed" matter, even for offensive linemen - skills like hand technique, foot placement, stance, balance, communication.
But even if we accept your hypothesis, despite the evidence to the contrary, and we accept that the more talented team will win a lot more, that still doesn't justify 0-5. Literally the only way that an 0-5 record should be expected is if Ohio State is more than 8 times better than Michigan (> 8:1 odds of beating Michigan in each game). Not even Vegas believes that.
We shouldn…
0-5 is the blip
We shouldn't expect that and we shouldn't settle for that.
You created a post claiming to provide a "Data-Driven" approach, and then you use horrible statistical assumptions and methods in your approach. I've shown you how your calculations are faulty, and you continue to dismiss data that doesn't agree with your predefined conclusion as "blips" or "lucky" or outliers.
That's horrible "Data-Driven" methodology, and I won't apologize for calling it out on a post with "Data-Driven" in the title.
The reality is that 0-5 is the statistical outlier, not the expected trend. We should expect better.
I'd also point out that Stanford and USC have been 50-50 since Harbaugh left, despite USC consistently having higher rated recruiting classes. There's your trend.
I agree with your latest point - that coaching does matter. In fact, I would argue that coaching and program building matter at least as much as raw talent, if not more. Furthermore, I would argue that in no reasonable comparison of Power 5 conference football teams, would talent dictate wins in the way you suggest.
I like to use the example of Stanford vs USC when Harbaugh was at Stanford, because it negates a lot of the preconceptions that criticism of an 0-5 record vs OSU is automatically an unfair criticism of Harbaugh.
Stanford under Harbaugh went 3-1 against USC under Pete Carroll. I think it's reasonable to suggest that Pete Carroll is not a significantly worse coach than Jim Harbaugh. Yet Stanford, with more stringent academic requirements, significantly lower rated recruits than USC, and a backup QB in the 2007 game, won 3 out of 4. In that situation, Harbaugh and his staff built a better program and coached the players more successfully than Carroll and his staff.
To now suggest that an 0-5 record against OSU is a reasonable expectation for Michigan flies in the face of actual evidence, including evidence of Harbaugh's own excellent work in the past. Harbaugh wasn't hired to recruit several spots lower than OSU and lose the Game every year. Harbaugh was hired because he had a demonstrated track record of winning against other top coaches and against more talented teams - even with stringent academic requirements and a clean program.
We have to accept that while Harbaugh has proven to be a good coach in the past and even in the present, he and his staff need to do better now and in the future. The lack of effective DT players is partially Harbaugh's responsibility, the lack of discipline and execution is partially Harbaugh's responsibility, and the 0-5 record is partially Harbaugh's responsibility. He has to own it, and we have to expect him to own it. Talking about a talent discrepancy as if it defines football destiny is simply not valid, statistically or otherwise.
I've said it before, football is notorious for small sample sizes. You have 5 data points and are arguing for a statistical correlation. Yet you could take a different set of 5 data points (say MSU vs OSU 2011-2015 or Stanford vs USC 2007-2011) and find an inverse correlation.
I have two problems with the general argument that more talented recruits somehow justifies the 0-5 record against OSU. First, 0-5 is statistically very different than any other record. If Michigan should lose 70% of the time because of talent, then an 0-5 record still only has a 16.8% probability. Winning at least one game has a 36% probability. Put another way - even if OSU's talent means it should win 70% of the time, any record other than 0-5 has a probability of 83.2%. Going 0-5 is the outlier - it should not happen. It should not be expected.
Second, the talent = wins argument devalues program building and coaching, yet we see the impact of program building and coaching, and the limited impact of talent, in every sport. There's plenty of contrary evidence of less talented teams beating more talented teams, and if it really all boils down to talent, then what's the point of hiring good coaches? Of having good schemes? Of complementary football?
We have to value program building and coaching, and we have to believe (and expect) that talent alone will not dictate what happens. We should expect better.
Your claim wasn't that we needed to be within 2-3 recruiting spots of Ohio State to be 50-50, your claim was that being more than 2-3 spots lower than Ohio State would result in 10-90.
If the recruiting difference between Michigan vs Ohio State truly results in a win probability of 10%, then the probability of an 0-5 record is about 59%. OK - Michigan is doing what's expected. However, that also implies, by your same standard, that the probability of MSU beating Michigan 7 out of 8 times is something like 0.0000009%.
By contrast, if recruiting rankings are not so powerful in determining the outcome of games, but are only an influence (say, 30% win probability instead of only 10%), the probability of going 0-5 against OSU is only 16.8% (and the probability of MSU winning 7 of 8 is still only 0.015%).
Football is notorious for having small sample sizes, but that works both ways in your argument. You make an implicit statistical argument - we shouldn't expect anything better than 0-5 because the probability of beating a team with more than 2-3 spots higher recruiting rankings is only 10%. Then you reject contrary evidence (Auburn, MSU, Stanford) as statistical outliers ("lucky events"). But if those are outliers, then Michigan's 0-5 performance is also an outlier - it shouldn't be happening. It shouldn't be expected.
I appreciate the analysis that went into this, but my question is whether Auburn has been consistently within 2-3 spots of Alabama in the 247 Sports composite rankings since 2013, when they beat Alabama 3 out of 7 times. Or whether Stanford was consistently within 2-3 spots of USC in those recruiting rankings from 2007-2010, when they (under Harbaugh) beat USC 3 out of 4 times. Or whether Michigan State was consistently within 2-3 spots of Michigan in those recruiting rankings from 2008-2015, when they won 7 out of 8.
If you were to conduct an analysis of recruiting rankings over time, is it really true that teams 2-3 spots higher in those rankings will win 90% of the time?
Here's the problem with your analysis. Time of possession for Ohio State:
* 2013: 33:21
* 2014: 25:41
* 2015: 29:54
* 2016: 28:47
* 2017: 28:43
* 2018: 24:36
* 2019: 23:35
Michigan's defense gave up more points in less time over the past 2 years. So saying that OSU had more possessions is not a good thing for the defense - it means that OSU was able to score more quickly, more easily, with less time. You've proven that OSU was able to cut through the defense like a "hotter" knife through butter - that's not a compliment to the strength of the butter.
That would be comforting if Michigan's defensive performance was comparable to "the best defenses" against Ohio State. The problem is that Michigan's defense gave up:
* More points than Florida Atlantic, Cincinnati, Indiana, Nebraska, Michigan State, Northwestern, Wisconsin, and Penn State, and gave up the same number of points as Rutgers.
* More yards than Florida Atlantic, Cincinnati, Indiana, Nebraska, Michigan State, Northwestern, Wisconsin, and Penn State.
* More yards per play than Florida Atlantic, Cincinnati, Indiana, Michigan State, Wisconsin, and Penn State, and the same number of yards per play as Maryland.
And that was with Ohio State possessing the ball for only 23:35, compared to their season average of 31:38
Except Parris Campbell is not on the 2019 Ohio State team. Ohio State's 2019 offense goes through JK Dobbins. Scheming to stop a player that's not on the team anymore, and failing to scheme to stop the key player on the current team, is not a positive adaptation.
No. How could anyone justify promoting Brown to head coach after the last 2 defensive performances against Ohio State?
I think it will take a special QB. That's a great equalizer. Where is Stanford without Andrew Luck? Where is Michigan State without Kirk Cousins or Connor Cook? Purdue without Drew Brees? Even when Purdue beat OSU last year, we've seen that David Blough is at least a backup-level NFL QB. Unfortunately, I don't think Shea Patterson was special enough.
While I agree that NIL rights will change things, I disagree that the playoff system didn't change anything. The playoff system not only provided more clarity about the national championship, it also focused more attention on the national championship as THE goal. Prior to that, getting to the Rose Bowl (or equivalent New Year's Day bowl) was considered the highest reasonable goal for any program other than Alabama, because the national championship was left to the unpredictable vagaries of "voters" and "media bias" and "scheduling."
Before the playoff system, recruiting between OSU and Michigan could be somewhat balanced, because neither program really had any reasonable claim to playing for a national championship, and both programs could provide equivalent exposure on New Year's Day. If anything, Michigan under Harbaugh might have been able to provide better exposure.
OSU winning the first playoff while Michigan went 5-7 under Hoke really established a clear distinction between the programs at a crucial time.
The reverse gambler's fallacy. The base rate fallacy. Simpson's paradox.
Big10 championship down to the wire. One goes to the playoff, and Pac12 champ is left out. Said Pac12 champ then destroys Big10 runner-up in the Rose Bowl, and Big10 champ destroyed in the playoff. No way the committee puts 2 Big10 teams in *at the cost of the Pac12 champ* this year.
No way the committee leaves out Pac12 champ in favor of 2nd Big10 team after what happened last year.
Questionable Big10 champ with close win over Big10 runner-up gets in over Pac12 champ, who promptly unpants said runner-up in Rose Bowl? I don't think the committee freezes out Pac12 champ 2 yrs in a row in favor of a 2nd Big10 team.
The reason Brady was suspended 4 games is because of the Patriots' opponent in their 5th game. Imagine the TV ratings...
And Brady wasn't asked to hand over his phone. He was asked to choose what texts were relevant and print them up for the investigators, and they would trust his judgment about what was relevant. He declined to do even that much.
Perhaps the emphasis is misplaced?
Also, why victims don't leave - perhaps that should influence how we talk about domestic violence?
http://www.ted.com/talks/leslie_morgan_steiner_why_domestic_violence_victims_don_t_leave/transcript?language=en
Don't write them off so easily... ;-)
There's a perception, accurate or not, that star athletes get special treatment by law enforcement. That perception then leads to a presumption - when law enforcement does take action against a star athlete, it is probably an under-reaction rather than an over-reaction. Hit-and-run turns into a traffic ticket, body slamming a security guard becomes a "Youthful Trainee" sentence, whipping a child until he bleeds becomes a no contest plea to a misdemeanor, etc.
Frank Clark's arrest occurs in this context. We can all advocate for innocent until proven guilty, waiting for the legal process to play out, making sure all the facts are in before jumping to conclusions, etc. However, I think we also have to recognize that the law isn't always blind when it comes to star athletes, and we can't always rely on the outcome of a legal process to completely define how we should view, or judge, a particular situation.
It's fair to hold our legal system to a standard that the accused are innocent until proven guilty. It's fair to hold reporters to a standard of framing their reports with the word, "alleged." But I also think it's fair for the public, including people who post on fan blogs, to view events like this within the context of how star athletes are treated by law enforcement in our society.
Well, they did their best.
http://youtu.be/w3UQwyKrTtI
I hope we're willing to recognize that "running up the score" is inappropriate even when it's done by someone (we want) on our side. However, just as some want to down play the "spear" because it was done by our players, I also suspect some are complaining about the last TD because we weren't the ones scoring it.
MSU went up 21-3 on Gardner's pick 6 - that wasn't the defense's fault. Except for the drive at the end of the first half, I thought the defense did a decent job with how little help it was getting from the offense - 2 net yards in the first half, only 10 minutes ToP, etc.
for most of the first half. Giving up a TD on the last drive of the 1st half was a killer. But without some life from the offense, the defense couldn't keep it up.
Michigan 1st half drives:
4 plays, 17 yards, punt, 2:01, 0-7
1 play, -9 yards, fumble, 0:07, 0-7
9 plays, 25 yards, punt, 5:46, 0-7
3 plays, -3l yards, punt, 1:43, 0-7
3 plays, 0 yards, FG, 0:48, 3-7
At that point, MSU had a 2:1 advantage in ToP. And Michigan had gained all of 2 net yards. Any defense would be hard pressed to keep up the fight in that situation.
as many times as Harbaugh would - which is every time someone complains about Dantonio running up the score.
http://blogs.mercurynews.com/collegesports/2009/11/14/stanford-beats-usc-the-two-point-conversion-attempt/
http://blogs.mercurynews.com/collegesports/2009/11/14/stanford-beats-usc-the-two-point-conversion-attempt/
http://blogs.mercurynews.com/collegesports/2009/11/14/stanford-beats-usc-the-two-point-conversion-attempt/
http://blogs.mercurynews.com/collegesports/2009/11/14/stanford-beats-usc-the-two-point-conversion-attempt/
http://blogs.mercurynews.com/collegesports/2009/11/14/stanford-beats-usc-the-two-point-conversion-attempt/
http://blogs.mercurynews.com/collegesports/2009/11/14/stanford-beats-usc-the-two-point-conversion-attempt/
http://blogs.mercurynews.com/collegesports/2009/11/14/stanford-beats-usc-the-two-point-conversion-attempt/
http://blogs.mercurynews.com/collegesports/2009/11/14/stanford-beats-usc-the-two-point-conversion-attempt/
http://blogs.mercurynews.com/collegesports/2009/11/14/stanford-beats-usc-the-two-point-conversion-attempt/
Administration matters. A good coach with a bad AD will have a difficult time being successful. A bad coach with a good AD can be replaced.
There's more to Michigan's reputation than athletics. How Shane Morris was handled does matter. Anyway, I'll post in Ace's thread - I had seen this one first (hence the edit).
The law doesn't require the public agency to retain email forever, so it's possible that the email was deleted prior to the FOIA request.
http://www.michigan.gov/documents/hal_mhc_rms_email_faq_161101_7.pdf
Yes, the regents probably could do with a little more sunshine. But realistically, if the other regents believed that anything they said could end up posted on MGoBlog, they would stop talking to Brian except in the formal meetings. The closed doors would just move somewhere else, and exclude Brian.
Also, Brian would probably be forced to start editing himself on MGoBlog to avoid creating legal liabilities for the University. As a regent, he would have a duty to the best interests of the University, not to us, and would have to honor things like FERPA and HIPAA, legal counsel's opinions, etc. Then we, the readers of MGoBlog, would stop trusting Brian as much because we would sense he's not telling us everything.
It's very tricky to try and be both an outside advocate and an inside policy maker. Usually, it doesn't work out very well for the person in the middle.
Maybe going for 2 right away is seen as being aggressive and trusting your offense to get the job done, and trusting your defense to get an extra stop if necessary? Feelingsball can always go in multiple directions.
You're being hypothetical, and we're not going to work in hypothetics.
Win or lose, I'll be looking for sustained discipline. I think that playing with emotion, but keeping those emotions under control, will be the most challenging part of the game this Saturday. Given all the distractions and emotions of the week, I could see emotions running high but losing discpline (penalties, miscues, etc.). I could also see emotions running low, and a flat performance or folding in the face of adversity.
Even if they lose, if they appear to keep their emotional edge but stay in control, that will be a good sign for the direction of the team.
Even if they win, if they appear to be undisciplined or flat, that will be a bad sign for the direction of the team.
If people were calling for Brandon to be fired by claiming he slept with a student assistant, or claiming he embezzled funds from the University - that would be a witch hunt. Calling for him to be fired for factual aspects of his job performance is not a witch hunt. You can debate whether his job performance merits being fired or not, but you can't dismiss those calling for his ouster by claiming they are conducting a "witch hunt."
Let's imagine 2 different drivers are pulled over for speeding by the same police officer on the same road going the same speed. One has a clean record, immediately takes responsibility and commits to ensuring it doesn't happen again. The other has a record of multiple traffic violations, reacts defensively to the officer, and insists that it was the passenger's responsibility to keep an eye on the speedometer.
While both drivers might receive speeding tickets, we can understand if the officer chooses to issue a warning to the first driver, or if the court applies a more lenient sentence to the first driver.
Context does matter, and it's unfair to the protesters to insist that they should always treat each event in isolation and ignore the context.
So Brady released a statement to the press last night without consulting his boss? Sounds like grounds for Brandon to fire him. /s
An alternative might be that Morris didn't have a concussion, but they're concerned about admitting that they didn't follow the concussion protocol at the appropriate time.
If they evaluated Morris after the game and found that he had no concussion, it would still cause some potential problems (including liability) if it was revealed that the evaluation didn't happen until after the game.
Still, there might have been a way to finesse the language to obfuscate the timing. So if he was evaluated after the game and found not to have a concussion, there probably would have been a way to get the good information out (no concussion) without being too specific on the bad information (evaluated after the game).
So, that's just long way of saying your explanation is most likely accurate.
Just answering your first question - they could announce a review of player safety protocols by a credible 3rd party or by a credible faculty group. Universities have structures and protocols for all sorts of situations that have ethical questions, ie research involving human subjects, etc. It doesn't have to be an extraordinary thing for a University to study itself.
"...we ran out a coach who had improved each year but not fast enough for us. He has, of course, become the head coach at another program, improved that team each year..."
Can we say the same for Hoke?
A Power5 conference school is struggling and hires a new coach:
After only 2 years, he's fired. The AD gets a lot of heat for firing a proven coach so quickly.
"This decision...was a result of not seeing the progress in some of the critical areas we would have hoped for after 2 years...I think the most critical assessment I had to make was, does investing another year present the likely possibility of substantial improvement? If I couldn't answer that affirmatively, which I ultimately didn't, I felt it was better to make the change now than wait more time."
The AD? Bob Bowlsby. The fired coach? Walt Harris. The school? Stanford. The replacement? Jim Harbaugh.
Didn't take the time to go through the entire history of Ball State football, but since 1971:
19 seasons in which Ball State was outscored by its opponents:
1971 (McClain)
1981-1984 (Wallace)
1985-1987, 1992, 1994 (Schudel)
1997-2002 (Lynch)
2003-2005 (Hoke)
10 point losses:
Seasons with four or more 10 point losses:
Didn't take the time to go through the entire history of San Diego State football, but since 1973:
20 seasons in which San Diego State was outscored by its opponents:
1978, 1980 (Gilbert)
1982-1983 (Scovil)
1987-1988 (Stolz)
1989, 1992 (Luginbill)
1994, 1997-1998, 2000-2001 (Tollner)
2002, 2004-2005 (Craft)
2006-2008 (Long)
2009 (Hoke)
10 point losses:
Seasons with four or more 10 point losses: