Sagarin B1G ratings: In BB, UM #1, in FB UM still >Ohio.

Submitted by michelin on November 25th, 2012 at 12:19 PM

According to the Sagarin ratings, which predict actual game outcomes:*

IN BB, UM is #1 in the B1G.  In the nation, UM is #3,  IND #4, and Ohio is not in the top 10.*

IN FB, on a neutral field UM would still be favored over Ohio (by 0.15 pts).   Why? UM’s nonconference opponents included the two teams now favored to be in the national title game.  By contrast, Ohio’s “marquee” nonconference game this year supposedly was Cal, which is now not even in the top 70 nationally.  Moreover, Ohio’s  other wins were often very narrow.  In fact, even for the UM game, they were only +2 pts after subtracting 3pts for home field (not even considering the injury to UM’s starting QB). 

What then should we make of Ohio’s claim that they could win the AP national title?  That claim should be laughable to any educated voter.  In fact, Ohio is not even ranked in the top 20 nationally by Sagarin (they are #24, whereas UM is #22).  Also, Ohio will not be tested in a competitive bowl or conference championship.  Indeed, in the latter, on a neutral field, they would be  favored by less than half a point vs WISC, 2.5 pts vs NW and they would be underdogs to both NEB and UM.

Interestingly, ND's delusions of grandeur also should be tempered.  Although human pollsters will no doubt put them #1, would ND actually be favored to beat all the other teams according unbiased Sagarin PREDICTOR ratings?  No, not at this point. Why?  ND had many narrow wins, even over marginal teams and teams expected to be huge challenges--like USC--turned out not to be so great.  Thus, by Sagarin's ratings, while ALA is #1 and Oregon #2, ND is only #3  Both ALA and OR both would be favored over ND by large margins on a neutral field.  In addition, unlike ALA but like Ohio, ND will not be tested in a conference championship game.


*I report only those ratings that predict actual game outcomes.  For BB I take the average of ELO and PREDICTOR ratings.   For FB, I report only PREDICTOR ratings (not the ELO-CHESS, which is used by the BCS but does not consider margin of victory or predict actual game outcomes).



November 25th, 2012 at 12:38 PM ^

you are pushing hard for a spot on TWIS.  Let it go man, this is an exercise in futility.  You are preaching to the choir when you say ND and OSU are not as good as their records because of X, Y, Z...but it doesn't change the most important number...W-L record.  

Rough year.  No amout of spin or rationalization will fix that.


November 25th, 2012 at 12:31 PM ^

It's their year.  Solid team on a karmic role.  Don't be surprised if Georgia upsets Bama and then Richt proceeds to get outcoached by Kelly in the championship game.

Even an ND upset of Bama wouldn't surprise me.  One where Bama outgains them like 450 yards to 120, but turns it over 5 times and ND walks away with a most unlikely 3 point win and national title we will never hear the end of.  It's been that kind of year.

At this point it wouldn't even surprise me if McCarron fell down the stairs and broke his leg the night before the BCS championship game.


November 25th, 2012 at 12:36 PM ^

Alabama is not going to run away from Notre Dame, as much as people want to think that, look at their schedules. Alabama's best win was against LSU and they should have lost that game. It will be a defensive slugfest and I for one hop Notre Dame wins. Hate on them all you want but they went to Oklahoma and won convincingly, beat Sparty on the road, and USC on the road.


November 25th, 2012 at 12:45 PM ^

A Bama blowout would not surprise me, but an ND miracle wouldn't either.  Bama's o-line will be far and away the best ND has faced this year, basically matching ND's biggest strength, their D-line.  Bama has an experience QB and I expect them to hit a few shots downfield.  

Despite their lunatic coach, LSU has a nasty, nasty defensive line and a better offensive line, skill position players and secondary than ND.  

The ND-Oklahoma game was tied in the 4th, so wasnt really convincing, Sparty couldn't score on Pioneer HS, let alone a top NCAA defense, and USC's defense was soft, swiss cheese.  As we saw, Bama is a different beast.  

Can ND win, sure, anything can happen.  But if both teams play to their ability, I don't see ND hanging.  But hell, they have proved me wrong all year.


November 25th, 2012 at 1:03 PM ^

Alabama didn't play a murderer's row, but show me when ND had to play three top 15 teams in a row this year.

Notre Dame played an average Oklahoma team that couldn't get out of their way.  USC played with a redshirt freshman QB and a team that was checked out emotionally.  And Sparty lost to Iowa at home.  Back in August, those teams looked a lot better on paper than the product they put on the field.

And ND was not going to win at LSU.  That's a fierce defense and a crazy atmosphere.  None of ND's wins were as impressive as that.  In fact, the same day Alabama was winning a game they apparently should've lost to a top 5 team, ND was squeaking by a sub-.500 Pitt team at home, with help from the refs. 

If Bama beats Georgia next week, they'll have two wins that are better than any victories on ND's schedule.  


November 25th, 2012 at 1:28 PM ^

I kind of agree.  The Alabama team that started the season seems to have slowed down a bit, given how they lost to A&M and nearly lost to LSU.  They beat up on an Auburn team that absolutely gave up on the season a couple of games ago, and have exactly two wins against current top-25 teams (LSU and UM).  Give Saban a month and he'll have this team ready, but there is a very real possibility they lose to Georgia next week.  And ND isn't going to be afraid of 'Bama either - they've played a whole season and beaten Oklahoma, Stanford, UM, USC, and an MSU team that has a solid defense.  That might not be as murderous a row in reality as it appears on paper, but it certainly is as good as Alabama or any other team's in the country.


November 25th, 2012 at 12:42 PM ^

we lost that one game by the equivalent of 2 points.  On the season, as a whole, however, considering our impossible schedule (and not even considering the injury to our starting QB), we rated higher than Ohio.   While we lost our two biggest nonconference games, both on the road, Ohio would also be predicted to lose these games by large margins.  Also, while we lost more games than Ohio, we won many games by large margins.

One may argue that the Sagarin results seem counterintuitive.  Nevertheless, the ratings predict future game outcomes better than just about anything else.


November 25th, 2012 at 12:33 PM ^

This basketball season has a chance to be historic. If you haven't watched them play yet. You need to start watching. This might be the best team Michigan has had in 20 years


November 25th, 2012 at 12:54 PM ^

you also need to consider margin of victory, as well as SOS, in order to predict future game outcomes.

According to your argument, a hypothetical 8-4 team in the SEC who upset ALA by a large margin on the road could not possibly be better than a 12-0 team, that needed overtime to win several games in the MAC.  That sounds a lot more ridiculous to me than do the Sagarin ratings.


November 25th, 2012 at 1:01 PM ^

What is the thing we did that's analogous to upsetting ALA by a large margin on the road? He's talking about Michigan, an 8-4 team that lost to 4 good teams. There's no big upset victory there. If we had spanked any of the teams we lost to, you'd be closer. But we didn't. We lost every game we played against good teams.


November 25th, 2012 at 12:44 PM ^

The basketball ranking is interesting for its predictive value, the football one not so much. The objective of the sport is still to win the game played, not win the majority of thousands of simulations.


November 25th, 2012 at 1:37 PM ^

The BB prediction seems more plausible on the surface.   The FB rating surpirsed me too.  However, it does make you ask deeper questions about why UM could have rated higher than Ohio.   It makes you look back much more closely at margins of victory, SOS and home field advantages.

As you rightly noted, these items have no bearing on the actual object of the game.  Also, you might argue that they have no bearing on our in-a-sense-"moral" evaluation of the goodness of our season  (although I am not sure we would all agree about the criteria for such evaluations).

The only point of the Sagarin ratings, however, is to make a prediction of future game outcomes.  eg How will UM do when playing another team on a neutral field in a bowl game? (We could ask the same question about Ohio, but unfortunately, due to their probation, we'll never know the answer).

The Sagarin ratings are desigened to optimized these predictions and often come very close to the Vegas point spreads.   Maybe some of us do not care about that.  But if we were placing very large bets, we should care.



November 25th, 2012 at 12:47 PM ^

Of our 4 losses. 2 were to the NC contenders. 2 (OSU and ND) we would have won if we didn't turn the ball over 9 times. And 1 was because of Denards injury and Devin wasn't ready.


November 25th, 2012 at 1:06 PM ^

Unless I am missing some attempt at sarcasm, I don't agree that Ohio is a serious NC contender, as he indicates.  ALA clearly is the serious contender.

I do agree with you that he must have meant that.  I am just trying to make that clear.


OMG Shirtless

November 25th, 2012 at 1:10 PM ^

Point 1 - Of our 4 losses. 2 were to the NC contenders [INSERT ND AND BAMA HERE].

Point 2 - 2 (OSU and ND) we would have won if we didn't turn the ball over 9 times. [OSU AND ND - OBVIOUSLY] 

Point 3 - And 1 was because of Denards injury and Devin wasn't ready. [NEBRASKA]



November 25th, 2012 at 1:01 PM ^

We only fell one spot in the AP football poll.  Now we have five weeks to heal and gear up for Georgia or Texas A&M.  In a New Year's bowl game. 

Plus BBall has beaten two tournament teams already, is ranked in the top four, with another on the docket for Tuesday. 

Overall, the Big Picture looks so much better now than it did at any time 2008-2010, which means more than we probably will admit it does. 

Dutch Ferbert

November 25th, 2012 at 1:12 PM ^ I think our interior O-Line has a higher WAR than ohio's D-Line. Thus, running the ball up the middle on 3rd/4th and short was the only choice.

Predictive stats are great until they are faced with the reality of the scoreboard.

Our team lost, but life goes on. Go Blue!



November 25th, 2012 at 2:05 PM ^

and I do appreciate your interest in the topic.  But I agree with the above criticism of your post.  Your sunday morning quarterbacking has little to do with the topic of the OP, which reports the results of  "predictive." ratings. 

Please excuse me for being repetitive but you don't seem to get it.  These ratings are based on "reality" just as much as are your selective recollections of the game.   However, the "reality" of these ratings includes not just the WL records but also SOS, victory margins, home field advantages etc which comprise the entire season.

Go Blue.

Dutch Ferbert

November 25th, 2012 at 2:19 PM ^

And I do get it. Do you really think leading my post with "I saw Money Ball" meant that I was serious about everything I said? Clearly stats are based on reality and more so than my selective recollections of the game. But my only recollection of the game that matters more than Sagarins stats is the final score of yesterday's game.

I actually enjoy reading most of statistical analysis on this blog (and I believe we have one of the most intelligent fan bases in football), but like everything, I take it all with a grain of salt.

Stats prediciting that we would win if we met OSU on a neutral field only dig the knife in a little deeper.

Go Blue!

Dutch Ferbert

November 25th, 2012 at 2:11 PM ^

I know that...and that was kind of my point.

WAR (like any statistic) is used by people to judge the value of a player, but athletes still need to go out on the field and make plays. Human error, broken ankles, crazy bounces, crappy refs or a freshman backup coming off the bench and having the game of his life can all negate the most precise stats.

Oh well, I guess my third post ever is flamebait. Yeah me.


November 25th, 2012 at 3:49 PM ^

OF course, human error, crazy refs and crappy bounces affect the actual outcomes but that does not entire negate the ability of stats to make predictions.

Also, there is a difference between merely reporting past stats and testing which stats actually make useful predictions.  If I find that a left-handed BB player this year has a .500 avg vs LH pitchers, clearly that reported stat does not predict he will get a hit every other time when facing any LH pitcher.   Yet, some stats are still useful.  If a batter with a  BA of .050  faces a picher with an ERA of 0.050 for a hundred consecutive at-bats,  then he will probably not succeed as often as when a .300 batter faces a picther with an ERA of 13.  

The Sagarin poll similarly determines how likely a it is that a team with a poor (or good WL) record will do when facing a good (or poor) opponent.  Through a vast body of game outcomes, it determines which stats predict future outcomes and it determines what weight--if any--on each one.  

Perhaps you do not feel that this prediction says anything about the quality of your team or its past season.   Fine.  But then tell me:  what other set of criteria do you propose to use to evaluate the quality of the season? 

For example, maybe you say: we should only consider the results of head-to-head matchups.  It follows that Texas A&M should replace ALA in the national title game.  But maybe somebody else feels we should also consider overall WL records.  It then follows that an undefeated MAC team should replace a one-loss ALA.  But somebody else then says we need to consider SOS.   And somebody else wants to consider home field (HF) advantages in eachgame.   So what do we do if all these  people disagree with one anothers' criteria (as actually occurred in BCS negotiations)?  How do we decide who is right and what weight to give to different factors eg HF vs SOS vs WL?

To resolve such disagreements, you need some criteria to judge who’s right and who’s wrong.   One way to do so is to make a prediction of how well a team will do in the future based on their stats (eg HF, SOS, WL) and then test the prediction.   The team predicted to win more often is judged to be the better team.   Maybe you don't agree with that.  But again, tell me a better way.  And tell me how we will know your way is better than someone else's? 


Dutch Ferbert

November 25th, 2012 at 4:39 PM ^

I don't think Sagarin's predictions serve no value. I'm sure they are right a large majority of the time.

But humans play the game, and there would be no joy watching sports if not for our human imperfections and the occasional superhuman effort.

To answer your question, the best way to determine the better team is to let them play. Ohio was the better team yesterday. A&M was better than Bama a few weeks ago. Maybe Sagarin is right and we would beat ohio if we played them again, but the only way to know if we would is to actually play the game.


November 25th, 2012 at 6:07 PM ^

Obviously, determining an outcome in the game is the whole point of playing.  It's the fun part.  That's why we create tournaments.  But to determine who gets in tournament, or who gets a BCS bid, a bowl invitation, or a national title shot, you need to consider more than the outcome of one game.   For instance, if one team beats another in OT by a single point at home and the other team has played a much tougher schedule, that does not prove that the home team is the better team in general  Head to head matchups tell us a lot but not the whole story.




November 25th, 2012 at 3:41 PM ^

One thing that I discovered, while pouring through the Sagarin numbers, was interesting to me. For purposes of comparison:

Michigan PREDICTOR rating is 82.31, and the average rating for the teams we lost two is 89.24, but the average rating for teams we beat is 64.52, though this is confounded some by an abysmally low rating for UMass. Strictly  within the conference, the average rating  for those that beat us is 82.37, and the average for the teams we beat is 69.44. 


November 25th, 2012 at 6:45 PM ^

Yet, these results would be reversed if we considered the fact that the number of possible points above us is much less than the number of possible points below.  For example,  there were about 18 possible points above us and the teams we lost to were 7/18 = 38% above us on average.  There were 82 posssible points below us and we the teams we beat where about 12/82=15 % below on average.

There are other ways you could also look at these stats.  But it would be interesting to see how UM's record compares with with other teams’ averages  vs winning or losing opponents.