Brian Fremeau projections for B1G

Submitted by ish on April 28th, 2014 at 5:49 PM

Link ($)…

Projects M to go 8-4 (5-3) and third in the East.  Projects an 8% chance of winning the B1G and a 4% chance of finishing with 1 or fewer losses.*

Calculations are based on his drive-based FEI.  He notes that last year we went 3-and-out on 38% of our possessions (92nd nationally) and had an explosive drive on only 12% of our possessions (73rd nationally). 

He projects OSU to win the East with a 30% chance of winning the B1G and a 31% chance of finishing with one or fewer losses.  MSU is 26/17%.  Over in the West, he likes Wisconsin (17/15%) and Iowa (13/16%).  But the West percentages show just how unbalanced the divisions are.

*Presented without comment as to what this record would mean for Hoke, etc.  Plenty of threads on that.


Everyone Murders

April 28th, 2014 at 5:59 PM ^

Not a criticism of the OP at all, but if I'm reading the post correctly these projections are based on last year's numbers.  And those numbers came from last year's players, who are not this year's players.  So color me skeptical (or confused).

Do we know how accurate these FEI-based predictions have been over the years?

Everyone Murders

April 28th, 2014 at 6:12 PM ^

That makes a bit more sense. I saw "Calculations are based on his drive-based FEI" and "last year we went 3-and-out on 38% of our possessions (92nd nationally) and had an explosive drive on only 12% of our possessions (73rd nationally)" in the original post, and thus wondered whether the Fremau predictions were made solely on those 2013 numbers. 

Does the formula take into account coaching changes and potential freshman impact players like Peppers?  Also, I'm still curious how accurate his forecasts have been over the years.

I Like Burgers

April 28th, 2014 at 6:22 PM ^

It does take into effect recruiting, so guys like Peppers factor into things.  But I don't think there are any good data to show the impact of a coordinator change so that doesn't factor in to things.  So depending how awesome Nuss is, and how incompetent Borges was, maybe you can bump the 8-4 up to 9-3 or 10-2.  Which for everyone's sanity's sake (mine included) I sure as hope to hell happens.

I Like Burgers

April 28th, 2014 at 6:02 PM ^

Yeah, I could definitely see 8-4 or on the positive side 9-3.  If Hoke loses 4 games or more this season, he would join Rich Rod and Harry Kipke as the only coaches in program history to post three straight seasons with 4 or more losses.  And neither of those gentlemen returned for another season.  So its likely 9-3 and better or bust for Hoke.

no joke its hoke

April 29th, 2014 at 3:06 AM ^

the excuse that bringing in a new staff is ugly only is true for programs that don't do it right. there are plenty of schools that get rid of coaches for whatever reason and brings in a coaches good or better. if this program could finally get it right, it would be way better than hoping that the staff turns it around next year. Bama,Ohio,PSU,MSU all did it right. sadly this program can't seem to do that.


April 28th, 2014 at 9:42 PM ^


Why the hell would we arbitrarily set a number cutoff before a single game is played? You really don't think losing close games while showing improvement across the board is no different than getting blown out in losses and scraping by in the wins? It doesn't make any sense.

Your criteria is arbitrary. Rich Rod didn't just have 4 loss season, he had 3, 5 and 7 wins. Harry Kipke had only 1, 4 1, 4 wins. Not to mention this was 1934-7, how the hell is that relevant? College football is just a little bit different today. And according to wikipedia he made it 4 seasons, so even by your arbitrary criteria, it doesn't mean Hoke needs to be necessarily fired.

Coaching changes are nothing to sneeze at. We need to really think this through because there will be a transition cost. There's no way to do that without playing the games and seeing what happens. Firing people just because we're angry and feel like someone needs to pay is a terrible idea. We got rid of what unnanimously was considered the biggest problem after last season, why don't we relax and give the new guy a chance before rushing to set meanless ultimatums for our head coach.


April 28th, 2014 at 10:03 PM ^

1997 and 2006 followed an 8-4 1996 and a 7-5 2005.

But...1996 included two end-of-season Top 5 road wins (Colorado and Ohio State) while 2005 included a Top 5 home win (PSU).  Further, all of those losses in those years could have been wins.  

Now, on the other hand, if we lose games like we did to Kansas State, then buckle up (for the rocket ship out of Ann Arbor).


April 28th, 2014 at 6:03 PM ^

Honestly, I can't see us going better than 8-4 this year. Also, I would hope that 8-4 means Hoke stays. Looking at the team and the schedule (three toughest opponents all on the road) it's tough to expect better. If meeting this year's expectations is going to cost him his job, then he should already be gone, as otherwise it's just another round of The Process. If Hoke is gone without 10 wins this year, we're all just wasting our time, since 10 wins isn't happening.


April 28th, 2014 at 6:13 PM ^

I mean, that may be true about 4 losses. But Rich Rod lost 6 games or more every year. That's not quite the same as losing 4 or more every year. And Kipke lost 4 four seasons in a row, not merely three, and they only played 8 games per year back then. The 4 losses thing is quite a stretch.

I Like Burgers

April 28th, 2014 at 6:19 PM ^

I don't think its a stretch.  The four losses thing is just a data point to illustrate that if he posted an 8-4 record this coming season, there's only two other coaches in program history that have had a worse stretch.  Rich Rod is clearly the worst, then you'd have Kipke who lost four while only playing eight game seasons, and then you'd have Hoke.  Every other coach in program history pulled out of the tailspin.


April 28th, 2014 at 6:28 PM ^

An 8-4 season would mean that last year's record of 7-6, his worst season, would be as good as or better than any season on the 7 final seasons of Rich Rod and Kipke. When your worst season is better than your comparables' best season, you are not actually comparable.

It's fine to want Hoke gone. But don't make up arbitrary numeric cut-off lines to support your case.

I Like Burgers

April 28th, 2014 at 7:19 PM ^

You're missing the point.  Its not a comparision.

The point is that by going 8-4, it would be the third worst three year stretch for a coach in program history.  I'm not comparing Hoke to any of those coaches.  I just named them because people would clearly wonder who the other two were (although one is a given). The four loss cutoff was just to illustrate what exactly qualifies as a bad three year stretch in program history. But if you want to fixate on the four loss part of my comment and how arbitrary or unfair of a comparision it is, then fine. To each his own.

Also, I'm not campaigning for Hoke to get fired.  I'd like to see him turn it around this season and have a stellar 2015.  He's still the best chance the program has at being strong again.  A coaching change is just going to take the team back a step.  However, if he has a bad 2014, there may be no choice but to reboot again.


April 28th, 2014 at 7:28 PM ^

You're making a 4 loss cutoff and defining that as bad, because conveniently if Hoke goes 8-4, he'll have hit that. Never mind that if he goes 8-4 (we'll presume he doesn't see the bowl game), that his abominable 3 year record will have been 23-15. And that's very similar to Kipke's 4-year record of 10-22. Or Rich Rod's 15-22.

Also, your 4 loss thing is factually wrong. Bump Elliot had three consecutive four loss seasons, but survived them (for a year). Hoke's hypothetical stretch is not the third-worst in Michigan's history. Bump Elliot was definitively worse for a decade.

I Like Burgers

April 28th, 2014 at 10:22 PM ^

Here's how I got here: I thought hey, they are projecting an 8-4 season and that would be the third season in a row with 4 losses or more.  And then I thought, hmmm...I wonder how many other coaches have had a stretch like that at Michigan?

So I looked it up.

And then I saw that only two other coaches have lost four or more games for three years in a row.  And I thought, hey, that's kind of crazy that in the 133 or whatever seasons of Michigan football only two coaches have had that many losses over a three year stretch.

And then I (apparently) made the FATAL flaw of trying to share what I learned with the board.  And now I have you, bluesalt, and some other dude shitting down my throat like I am somehow making things up and twisting facts or something.

So I guess fuck me then for trying to share something I found interesting.


April 28th, 2014 at 10:13 PM ^



1959 4-5
1960 5-4
1961 6-3
1962 2-7
1963 3-4-2

In the best of those years, Bump went .666, exactly what 8-4 would be.……

Bump returned in 1964 to lead Michigan to a B1G Championship 9-1 record and a 34-7 victory over Oregon State in the Rose Bowl.


April 28th, 2014 at 11:14 PM ^

It is just five years of the worst two decades in MI football, but it does show that we can bounce back as all great programs have.  That's why they're great.  Look at how long USC was mediocre after McKay and we all know Bama was not reall Bama between the Bear and Saban. Yes Stallings got a NC but they out violated Tressel during that period.  Hell if Oosterbaan had not inherited a couple of classes from The Great One I don't know what kind of overall record he would have had. He had a great two years to build on, but chose to recruit w/o ever leaving AA.             Bump also had a good record in '68, the year prior to Bo. Two out of ten is not good; this I understand.  However, unlike Carr, he left Bo with more future AAs than any other five year group in MI history. Aside from the debacle that was Missouri in '69, the pendulum had swung.            But why I like this post is because as much as Bump, or Chump, called so my my brother-in-law who played for him is that no matter how much he loved MI, even with the great players he had, he truly only had one great year.  Hoke seems to hold Michigan in the same regard and he recruits very similar to the way Bump did.  However, it took a great coach to get the talent out of the players and Brady hasn't been able to instill this into his team.  Outrecruiting everyone in the MAC or Mtn West might mean something, but when your major rival is recruiting even a tad better with a far better coach and your second greatest rival is beating you with far less talent it does, indeed, point to a lack of something on the part of the head man.  He can't blame his coordinators this year. On one side he has a multiple BCS NC coach and on the other he has a winning Super Bowl coach.   We all know what's he lacking and I've wrote about it often.  I always hold out and support M coaches, but when you replace someone who actually had to build from fricking scratch and had proven he based on wins and losses, despite not having talent, and one recognized as being among the best in the game,  you had damn better do so with someone who has won under the Big Tent.   


April 29th, 2014 at 11:34 AM ^

The point is that by going 8-4, it would be the third worst three year stretch for a coach in program history.


It just isn't true is all...


Hoke would have won 60.5% of his games over the three years (if they go 8-4 this year).


Post Yost, the following coaches all had three season stretches worse -


Rich Rod -

40.5% over three years


Kipke -

37.5% over three years (1935-37)


Bump Elliot -

44.0% over three years (1960-62)

46.7% over three years (1965-67)

48.1% over three years (1960-62)

53.8% over three years (1962-64)

55.5% over three years (1959-61)

59.3% over three years (1963-65)


Oosterbaan -

55.5% over three years (1951-53)

56% over three years (1956-58)


April 28th, 2014 at 7:43 PM ^

What a horribly misleading stat that was to use. If you wouldn't have posted this I might have actually thought it had some merit.

I hate people seizing on meaningless stats to "prove a point" like Hoke has to lose less than 4 games or he is likely fired. Self-serving crap.


April 28th, 2014 at 6:26 PM ^

Brian Fremeau contributes statistical analysis (along with a few other folks) to a blog called "Football Outsiders". The basics of FEI are here - LINK - and somewhere else on the site, it actually explains how these ratings grew out of an attempt forecast the NCAA basketball tournament while Fremeau was a student at Notre Dame, if I remember correctly. 


April 28th, 2014 at 8:47 PM ^

These guys are an analytical football prediction group. They are the Baseball Prospectus of football. Their system does a decent job predicting the NFL but with college football it is problematic given the high player turnover and uncertainly between seasons in college versus the NFL.

3% chance of one loss or fewer, what does this actually mean? It is an arbitrary number spit out of a formula which is slightly more accurate than throwing darts at a wall.

My non-analytical prediction says Michigan will win all home games and lose one or two on the road.


April 28th, 2014 at 6:09 PM ^

Yes our toughest three games are on the road but all of the other games suck. A depleted PSU at night at home, App State, a 5-7 Utah team, Miami of Ohio, Indiana, at Rutgers etc.


April 28th, 2014 at 10:33 PM ^

Utah is no gimme, but Appy St was far worse than Connecticut last year.



Rank Team Record SOS
34 Utah 5-7 3
44 Michigan 7-6 40
121 Connecticut 3-9 69
175 Appalachian St 4-8 183



Rank Team Record SOS
21 Michigan 9-4 37
44 Appalachian St 13-2 123

The 44th team beating the 21st team isn't all that unusual of an event. Sagarin would have favored us at home by about 10 points.

Arizona Blue

April 28th, 2014 at 6:09 PM ^

If the last two years have taught us anything its that all the pre-season noise is simply noise. I am pretty sure no one predicted the catastrophe that was last year and the revelation that was 2011.

Its a roller coaster fellas...just hold up your hands and hope the ride is fun