Minnesota Football Game Predictions: Iowa Hawkeyes

Time to blitz these HATE BREWS.

Models

During the Gopher bye week, the Battletoads continued their ascension to the top of the M&B ranks and overtook the entire state of Mississippi in the process. After narrowly escaping against the Holgo rage, they remain computer champions. Georgia? Think about the choices you've made in life that brought you to this point.

Rank	Team	BayesianProb	Delta
1	TCU	.932	0
2	Mississippi	.929	0
3	Mississippi State	.923	0
4	Baylor	.911	6
5	Oklahoma	.910	2
6	Alabama	.902	-2
7	Oregon	.893	2
8	Michigan State	.893	-3
9	Auburn	.893	2
10	Ohio State	.887	3
11	Kansas State	.885	1
12	LSU	.882	-4
13	Nebraska	.856	2
14	Florida State	.851	4
15	West Virginia	.851	2
16	Arkansas	.846	-2
17	USC	.834	7
18	Marshall	.833	2
19	Notre Dame	.821	3
20	Wisconsin	.820	6
21	Memphis	.814	0
22	Clemson	.800	3
23	Georgia	.800	-17
24	Utah	.793	-1
25	Arizona State	.778	5
34	Iowa	.760	13
64	Minnesota	.616	1

The loss to Illinois dropped the Gophers 20 slots in the ranks, while drubbing a walking hyperbole vocoder gave the Hawkeyes a huge boost. Due to the way the model handles post-game adjustments, I suspect there's a bit of overreach on both ends of that spectrum. Yet another thing to examine in the offseason is a more efficient adjustment mechanism.

On to the HATE projections:

	Flex Model
Gopher Win Probability	33.7%
Predicted Margin of Victory	-6.77
Current Vegas Spread	PICKEM
Current Vegas Over/Under	44.0

What? That doesn't seem right...

Much, much better.

This game is where I wish I had a home field adjustment (HFA) built into the model. Briefly, determining whether the game is played at home versus a neutral site is a significant co-variate for points scored (it is, by the way) is simple in a probabilistic model. However, my model is only partially stochastic and instead based more on a purely mathematical relationship.

Furthermore, I suspect HFA has similar properties to "Park Factors" in sabermetrics, meaning stadiums may have potentially differing degrees of expected points added (or lost) and those HFAs may also vary by other random effects, such as weather, rivalry, etc. That quickly becomes a more involved and complicated exercise, one which I'd love to explore but for the sake of evaluating performance of this iteration of M&B I chose to put off. I like the (mathematical) simplicity of this model and didn't want to corrupt it with endogenous factors like HFA.

In English, I think the TCF Bank Stadium HFA to Minnesota may be different in November, in 35 degree weather, against a primary rival than it would in late August against a scrub. That may bring the expected margin and probability closer to a PICK'EM or a 1 point Gopher win, as the F/+ projections suggest*.

No sims for the week, since it really doesn't matter right now. Beat Iowa, then we'll talk the rest of the year.

Visit Team	Home Team	Home WP%	Expected Margin
Bowling Green	Akron	49.1%	-0.36
Colorado	Arizona	80.1%	13.92
Notre Dame	Arizona St.	43.3%	-2.69
South Ala.	Arkansas St.	61.1%	4.5
Texas A&M	Auburn	73.9%	10.4
Northern Ill.	Ball St.	54.7%	1.9
Louisville	Boston College	36%	-5.74
Hawaii	Colorado St.	74.9%	10.95
Virginia	Florida St.	75.4%	11.21
San Jose St.	Fresno St.	34.4%	-6.44
Tulane	Houston	87.3%	19.32
Penn St.	Indiana	25.8%	-10.58
Iowa St.	Kansas	32.4%	-7.33
Toledo	Kent St.	27.6%	-9.64
Georgia	Kentucky	33%	-7.09
Alabama	LSU	44.9%	-2.07
Ohio St.	Michigan St.	51.5%	0.61
Iowa	Minnesota	33.7%	-6.77
Boise St.	New Mexico	20.5%	-13.55
La.-Lafayette	New Mexico St.	15.8%	-16.71
Georgia Tech	North Carolina St.	26.5%	-10.22
Fla. Atlantic	North Texas	41%	-3.64
Michigan	Northwestern	30%	-8.47
Buffalo	Ohio	48.4%	-0.63
Baylor	Oklahoma	49.7%	-0.13
Washington St.	Oregon St.	49.5%	-0.19
Wisconsin	Purdue	21.2%	-13.1
UTSA	Rice	72.6%	9.76
Idaho	San Diego St.	73.7%	10.29
Marshall	Southern Miss.	9.6%	-22.45
Duke	Syracuse	27.9%	-9.48
Kansas St.	TCU	63.9%	5.71
Memphis	Temple	32.6%	-7.24
West Virginia	Texas	33.2%	-6.99
Ga. Southern	Texas St.	14.9%	-17.44
Georgia St.	Troy	60.4%	4.22
SMU	Tulsa	79.2%	13.39
Louisiana Tech	UAB	29.4%	-8.76
Air Force	UNLV	26.4%	-10.25
Oregon	Utah	31.4%	-7.82
Florida	Vanderbilt	18.2%	-15.04
Clemson	Wake Forest	14.7%	-17.58
UCLA	Washington	42.3%	-3.1
UTEP	Western Ky.	32.8%	-7.18
Utah St.	Wyoming	36.4%	-5.58

* I'm 99% sure Bill Connelly uses a HFA and traditional regression methods for his F/+ projections, which explains how Minnesota is projected to win despite ranking 4 spots lower in F/+.

Bottles

After the debacle in Champaign-Urbana, this section should be nothing but grain alcohol. We won't go that far this week... yet.

What is the most appropriate bottle for HATEWEEK? Should it be a 1.75 of Hawkeye? Nah, we can't possibly usurp Vodka Samm's raison d'etre and besides, plastic handles don't count.

Bitter Neighbor Black IPA? That was last year's pick, and the results weren't too kind.

Iron City? I mean, Hayden Fry did pilfer other things from Pittsburgh...

Doppleganger? There are at least four different microbrews that make a beer with that title, which is too common for my tastes.

Templeton R-hahahahhahah nope.

No, as monsieurs Vint and Jacobi pointed out in the latest BHGPodcast, when Iowa and Minnesota look at each other across the field, they really see their reflection staring back at them. We've slowly but assuredly become the mirror image program, down to the same warts and complaints from the fanbase regarding just about everything. Appropriately enough, this week's HATE brew edition is Mirror Pond EPA, from Deschutes.

Drink up -- you're going to need that warmth for this bitterly cold tailgate.