Video for Mike and Mike why Michigan deserve to be in CFP Playoff

...but with an ever changing CFP committee and no public statements by individual members, this is basically a shot in the dark.  They're no better than Mike Valenti putting us #7 or #8 in his BS rankings.

Do you follow 538? They actually gave that recent event a far higher likelihood than anyone else, and did not write the possibility of it off at all. That said, I'm a huge fan of Nate Silver--but I think the expansion to the new 538 has been lackluster.

538 could, of course, be right in predicting how the selection committee will behave, but the numbers seem pretty far off from what I would expect. I think there's a decent chance there's just not enough data on what the CFP selection committee will do to make any kind of reasonable model (they use the AP for a lot of the examples of how people vote, but we've seen some definite differences between the selection committee and AP voter behavior).

One potential problem with the model, specifically related to Michigan is with head-to-head comparisons. When describing their Elo portion of the model (which represents the committee's decision making) they don't mention weighting head-to-head games at all (hopefully there's some in the model, but it's at a minimum not a factor they consider as highly as others, and it may be hard to do in a large simulation since you're really comparing all the teams collectively, rather than any two individually). Head-to-head games is really Michigan's strength here--if the committee doesn't care about them, it's easy to see Michigan being jumped from behind. If they're important to the committee, then it's hard to jump Michigan since the main potential teams have lost to Michigan.

If you predict a thing to be 35% likely, and everyone else predicts said thing to be <10% likely, and said thing happens...

That prediction wasn't "wrong". 35% likely means just that - it will happen sometimes (probably 35% of the time, in fact). Giving 6.5 points to Michigan against wasn't "wrong" - Michigan outperformed that, but that's expected variation. It's why you play the game. 

But I do think 538 is underrating Michigan's chance here - they are going off of insufficient data and not weighting the particulars of this year highly enough. 

I agree they are trying to get clicks(That is the point of making a website, right? You want to make money), however, they have consistently been one of the best sites in predicting future results in political races and sporting events. Nate Silver is known as one of the smartest minds out there in terms of probability and statistical analysis. 538 as a website has lost a lot of respectability due to the most recent election, but if you listen to Nate Silver, he explains that all of his numbers were actually accurate, but his team decided to use "common sense" in giving the president-elect a lower chance of winning than the numbers suggested. Before then, 538 was probably the most accurate in predicting previous elections(as well as sporting events). Definitely not a random website

Yes but this many blind squirrels finding a nut at the same time seems statistically unlikely.

At the same time, the whole meaning of "conference champion" has changed now that leagues have unbalanced divisions and the title comes down to a single game in December.

Whaaa???  Are you serious?  That doesn't seem right - they must be planning something.  They can't seriously think they'd beat us on a nuetral field...

That's fine for national titles, but it doesn't fix the Big Ten's problem.  Divisions worked well enough when they were 5 teams, OK when it was 6, but with 7 teams in each division its just not a good way of figuring out the two best and determining a champion.  Especially when the divisions get imbalanced.  Has the Big Ten had either of its top two teams in its championship the last two seasons?

I mean, look at the championship game matchups this weekend.  Four of the teams in the game are one of the top two teams?  That's dumb.

I don't see why the winners of these conference title games deserve an automatic playoff bid.  If an 8-4 team pulls the upset, it doesn't suddenly become an elite team, worthy of national-title contention.  College football has never functioned that way.  It's always valued the entire season.

Perhaps conferences should rank its teams not on record but rather SOS, quality wins, quality losses, etc, in the same manner as the committee, AP, coaches rank teams nationally?  This is just an absurd thought but whatever it's all a crapshoot anyway.