It all comes down to this. You ready for copious amounts of models and bottles on the most thankful of weeks?
Hello Church Oregon, good to see you again (now go schedule someone in the non-con). WPS bullies their way into the Top 6, along with Sparty? Bama called off the dogs ahead of their Iron Bowl showdown, so their drop is a function of the SEC's time honored tradition of scheduling FCS opponents in November, because ESS-EEE-SEE is so dayam tuff PAWWWWLLLLL.
|Rank||Team||Bayesian Prob||Last Week Delta|
I'd like to take a moment and appreciate what M&B is telling us. I had many reasons for choosing a new methodology this season, not the least of which was a desire for simpler and more concise insights. It's provided that, and then some.
Fundamentally, the M&B method rates how well a team is playing right now, with an eye looking back to all previous results. It is not a rating of a team's body of work from a win-loss perspective (like a poll, the CFP committee or Brian Fremeau's Degree of Difficulty metric does), nor does it simulate the entire season based upon the most up to date information (like the KRACH system for hockey does). I think the latter is what a lot of folks perceive a computer based rating system to be, and while I don't believe the rankings would change significantly if I implemented a forward looking simulated view, it probably would address the question of "who are the best teams in the country" better than the current method.
That said, this is a predictive model, and knowing everything I do from my time following the sport so intensely, how well you're playing now is most certainly a factor in how you'll play in the immediate future. Fans call it momentum, statisticians call it serial or autocorrelation. It's not foolproof -- no forecasting methods are foolproof -- but I do believe the model generalizes well and gives a realistic set of predictions/team ratings based upon every bit of information known to date.
Models and Bottles Refreshers!
Models and Bottles Refreshers!
Another insight? M&B has provided what I feel is a more representative set of projected margins than Vegas, as the sportsbooks simply have whiffed on more than a few lines. Where Vegas saw Michigan and Nebraska as double-digit favorites, M&B thought it was closer to a 1 point contest. M&B though Purdue would be closer, and that SJSU would still be a 17 point blowout. The model was closer on the Ohio State final margin than Vegas, with the only real misses against Northwestern, Illinois and Iowa (PELLET WAKE'D).
Lastly, the M&B rating is the posterior probability that a team will beat a perfectly average foe. What's a perfectly average foe, other than the tautology of being .500? Teams with a rating near .500 are roughly ranked in the 90s. FIU, Wake Forest, Arkansas State and Army are the teams closest to .500 in the M&B ratings, and they're all really bad. Anecdotally, Wake beat Army by 3. Wake beat Virginia Tech, a team with a rating of .608, by 3 points in double OT. In context of the model, Virginia Tech would be expected to beat a team like Wake Forest 60.8% percent of the time.
tl;dr. On to #HATEWEEK!
|Gopher Win Probability||38.6%|
|Predicted Margin of Victory||-4.64|
|Current Vegas Spread||+14|
|Current Vegas Over/Under||50|
Yep, once again M&B is more optimistic about the Gophers' chances than the bookies. This projection is actually even more favorable than the last HATEWEEK, where the models spotted the Hawkeyes as a touchdown favorite.
No outright boat race predicted. The details of how Wisconsin's win probability is 61.4% and the projected margin is 5 points relate to the deterministic component of the model, and are just that -- details. This will be a close, hard fought game between bitter rivals with everything on the line.
As for the rest of Hate Week...
|Visit Team||Home Team||Home WP%||Expected Margin|
|Colorado St.||Air Force||25.4%||-10.75|
|New Mexico St.||Arkansas St.||80.2%||13.99|
|Utah St.||Boise St.||53.5%||1.41|
|Ball St.||Bowling Green||55.1%||2.06|
|Old Dominion||Fla. Atlantic||43.9%||-2.46|
|West Virginia||Iowa St.||19%||-14.52|
|North Carolina St.||North Carolina||58.6%||3.47|
|Mississippi St.||Ole Miss||38%||-4.89|
|Michigan St.||Penn St.||16.4%||-16.31|
|San Jose St.||San Diego St.||84.1%||16.63|
|Notre Dame||Southern California||50.7%||0.27|
|Northern Ill.||Western Mich.||61.4%||4.66|
I'm half inclined to morph the bottles section into a choose your own adventure book, but in this case, your own libation. My reasoning? It's Thanksgiving, I'm thankful the Gophers are playing in the game and the week I've been waiting my whole life for, and I won't begrudge anyone for choosing quantity over quality. Thanksgiving is a holiday all about the bountiful splendor, which includes booze and football.
Most of us have never experienced this, so I don't know what the protocol is. Do we crush cheap domestics all game out of nerves and switch to the more appropriate post-game future hangover? Should we savor the moment and stem the tide with a higher quality offering to remain more conscious of the moment? Preeminent blackout strike against the the threat of disappointment with handles of rail gin? There are no good answers here, mostly known and unknown unknowns.
Previous selections for HATEWEEK include Great Lakes Nosferatu, an imperial red ale, and local product Bad Axe from Big Wood Brewery. I'm not a fan of repeat selections unless warranted, and this is not one of those situations. This week calls for something special, because that's what this season and this HATEWEEK is. Different. Extraordinary. Meaningful.
This is the Macallan 50 of HATEWEEKS, the rarest of whisky that will be experienced once, if at all, in one's life. Never again may we have the chance to say beating Wisconsin in the final week of the season will put us in the Big Ten title picture for the first time in many, many decades. If it happens, we have recent history. If it doesn't, well this game still happened, and the scenario won't be as foreign even if the stakes are the same. 50 year old scotch is reserved for those few pivotal moments that encapsulate the significant weight and scope of the moment, even if the moment is simply drinking from a $38,000 bottle. Insanely expensive? You bet. Worth it? Your mileage may vary. Appreciation of the fact it's even happening? Never in doubt.