B1G Championship Game Predictions

Submitted by The Mathlete on

Mid-Week metrics will be coming…well, mid-week.

Since we are within four weeks of the end of the regular, running some end of simulations start to make sense. Below are my projections based on my PAN ratings through last weekends games, run through 20,000 simulations and pushed through the Big Ten’s mildly confusing tiebreaking rules.

Bo Division

Odds of an outright win or two-team tie with straight ahead tiebreakers:

Michigan St: 58.3%

Nebraska: 29.0%

Michigan: 6.6%

Three Way Tie: 6.0%

Iowa: 0.1%

4 Way Tie: 0.1%

3 or 4 way ties mostly split between Nebraska and Michigan with a couple going Iowa’s way. Any 3+ team ties go against Sparty. The only judgment call I made was assuming in the event of a Michigan/Nebraska/Michigan St tie that went to the BCS computers I assumed that Michigan St would be out due to the ND loss and Michigan would lead Nebraska thanks to the head to head. Doesn’t change much, would swing a couple scenarios away from Michigan if they weren’t ahead.

In the end, Michigan’s odds are about 9.5% to reach the Championship game.

Update: MGoUser But I aint a Crip though had a good comment about Michigan's odds if they win out or go 3-1. Winning out gives Michigan a 37.5% chance of winning the division and going to the championship game. A loss within the division is a killer, odds reduce to 0.3% with a loss against Nebraska and 4.0% if they lose versus Iowa. An Illinois loss drops but winning the rest puts the odds at 21% and winning the next three but dropping to Ohio still means a 17% chance at the inaugural championship game.

Woody Division

Penn St: 42.9%

Wisconsin 41.6%

Ohio: 6.9%

Three Way Tie: 6.3%

Four Way Tie: 2.1%

Five Way Tie: 0.2%

Purdue: 0.02%

Despite a two game lead over Wisconsin, Penn St and Wisconsin are about even odds to win the division. Thanks to both of their losses coming out of division, Ohio wins almost any tiebreaker scenario they are in. There are a couple ties that the Buckeyes missed out on and Wisconsin and Purdue! split those.

If Michigan Makes It

Odds on who they would face:

Wisconsin: 48.6%

Penn St: 41.2%

Ohio: 8.8%

Purdue: 1.4%

 

Odds Used for Each Game

11/5

Minnesota @ Michigan St: 99% MSU

NW @ Nebraska: 99% Nebraska

Indiana @ Ohio: 99% Ohio

Purdue @ Wisconsin: 99% Wisconsin

Michigan @ Iowa: 67% Michigan

What a crappy set of games this weekend

11/12

Wisconsin @ Minnesota: 99% Wisconsin

Nebraska @ Penn St: 52% Nebraska

Michigan @ Illinois: 55% Michigan

Michigan St @ Iowa: 71% Michigan St

Ohio @ Purdue: 61% Ohio

11/19

Minnesota @ Northwestern: 77% Northwestern

Indiana @ Michigan St: 99% Michigan St

Nebraska @ Michigan: 56% Michigan

Penn St @ Ohio: 57% Ohio

Wisconsin @ Illinois: 72% Wisconsin

Iowa @ Purdue: 54% Purdue

11/26

Iowa @ Nebraska: 87% Nebraska

Michigan St @ Northwestern: 89% Michigan St

Purdue @ Indiana: 81% Purdue

Ohio @ Michigan: 70% Michigan

Illinois @ Minnesota: 89% Illinois

Penn St @ Wisconsin: 85% Wisconsin

Comments

Jivas

October 31st, 2011 at 2:14 PM ^

The computer models - PAN, Sagarin, FEI, etc. - don't know that Ohio played the first 5 games with Joe Bauserman and without Mike Adams, and that their team is fundamentally different at this time.  Therefore, I think Ohio is systematically underrated in these rankings.  On a neutral field, I'd think we're about 50-50 against Ohio; with the game at the Big House, I'd put our win likelihood at 60%.

Mathlete: I know the sample size of games with Braxton Miller as the QB (and post Tatgate suspensions) is small, but do you have the ability to make one-off adjustments to your system to test "what if" scenarios?  For example, if you implemented an assumption of a structural break in Ohio's ability starting with their sixth game?  Looking at their schedule it may not make too much difference - those close losses to Michigan State and at Nebraska came prior to the "break" and are impressive in hindsight - but it might be interesting to see.  At the least, those early games against Toledo and Miami (YTM) may need to be heavily discounted.

The Mathlete

October 31st, 2011 at 3:02 PM ^

Don't have a formal way of doing it but doing a quick check shows that Ohio was -6 on offense through week 5 and +0 since then. A six point swing is worth about 15-20% increase on every game which would swing all the odds, will try and add this in next week. It would also move Michigan from a 7 point favorite to a 1 point favorite.

BursleysFinest

October 31st, 2011 at 2:18 PM ^

 

  Nice analysis, The Big Ten is so weird this year, a whole lot of parity and/or suck (The Gophers/Hoosiers, etc) and no  truly elite teams,   I can't see Penn State winning out though, and as much as my heart wants to say Wiscy - UM....my head says 

Badgers - Huskers rematch, and the Badgers winning again

But I aint a C…

October 31st, 2011 at 2:34 PM ^

This is interesting (as always), but M's percentage looks heavily influenced by scenarios it loses 2+ more games.  More interesting: What are M's chances if they win out or drop only one more?  (Back of the envelope with your numbers says Sparty wins out 62% of simulations, so it's gotta be south of 38%.)

Ziff72

October 31st, 2011 at 3:03 PM ^

Those % are way too high.  99% should be reserved for games like Alabama @ home vs Ole Miss.

Northwestern vs an Offensively challenged Nebraska in a classic trap game is nowhere near 99%.

Many more like that.

One Inch Woody…

October 31st, 2011 at 3:57 PM ^

You should check the Massey predictions for Michigan and for other teams. For some reason, it has accurately predicted most of the margins of victory and records of most teams that I've looked at. He has Michigan at number 10 in the nation as: 75% favorites against Iowa (M 28 - I 21), 69% favorites against Illinois (M 24 - I 17), 53% favorites against Nebraska (M 28 - N 26), and 72% favorties against OSU (M 24 - O 17). Meanwhile, Michigan State is also heavily favored to win its last 4 games.

Miraculously, it predicted that Iowa v Minnesota would be a really close game, that Michigan State would trash tornado Michigan, that Michigan would beat Notre Dame, and that Illinois would lose to Penn State and Ohio State. If all of these predictions were truly made pre-game, then I'm simply astounded by how on-target they are.

Link:

http://masseyratings.com/team.php?s=107811&t=9031

Wolverman

October 31st, 2011 at 4:57 PM ^

 MSU will lose in Iowa or at Northwestern.

Msu is a different team on the road. Northwestern is good enough to score on MSU and their defense is ... (/gulp) good enough to stop MSU's offense in their house. I watched the MSU Iowa game last year ( granted Iowa lost some defensive players) and once again i think Iowa wins this game. Their offense is good enough to put points up against MSU and their defense plays well enough to stop MSU's offense in Iowa City. They are good against the run but have problems in the secondary... luckily cousins plays like garbage on the road.

Don

October 31st, 2011 at 8:17 PM ^

This is true for virtually all college teams; the real sign of a dominant squad is the ability to go on the road into hostile environments and still perform well. In our two road games, we played well in two out of eight total quarters; the other six were filled with turnovers, inconsistent defense, and an inability to run the ball out of the RB position. All three of Iowa's losses are on the road, and we should expect to see a dramatically different level of performance from the Hawkeyes next Saturday.

Roy G. Biv

October 31st, 2011 at 9:55 PM ^

I think M's chance of making Indy are negligible since Sparty gets to coast in from this point.  If they don't win out, some furniture will certainly need to be sacrificed in the streets.  They have a cake schedule remaining and obviously own the tie-breaker against the good guys.

I love the 70% likelihood of beating Tat U, but the last several years have taught me healthy skepticism.  God it would be nice to take their cheating arses behind the woodshed and leave their soulless husks of a fanbase demoralized.

jmblue

November 1st, 2011 at 11:30 AM ^

I'm surprised we have a better chance against Iowa than Illinois.  I feel like it's the reverse.  Did the loss to Minnesota dramatically change Iowa's numbers?

UofMedic

November 1st, 2011 at 4:20 PM ^

Perhaps I'm just missing something, but it seems a bit odd that Michigan could be favored to win out of the remaining four games, but be behind Nebraska in the odds in the division? Wouldn't it be MSU then UM?