For the last few years, I've blatanly stolen Seth's idea to use advanced metrics both to fill out my Bowl Pick'Em and to decide on which games to watch. Two years ago, using this approach got me 69% correct picks in my pool, but last year things were a bit rougher - an FEI-based pick'em got 54% correct, while a Sagarin PREDICTOR based one got 57%. When something doesn't work, throw more data it. So I put together a more elaborate spreadsheet (available here) that presents picks from several different advanced metrics: FEI, Colley, Massey (Power), and Sagarin (new, improved GOLDEN MEAN).
The methodology is straightforward - I compared all the teams using these metrics, and using the difference between them picked the winners and the confidence in the picks. That is, a huge difference in the ratings of the teams suggests a lock, a difference of zero is a push. In addition to looking at these metrics individually, I also put together a composite score by standardizing all the values and averaging them together. The list automatically sorts based on the system you use, with locks at the top and coin-flips (and presumably more exciting games) at the bottom. Interestingly, the four different system present three different potential national title winners, but none of those include OSU, so take some small pleasure in that.
Here is the table of composite picks:
|Rose Bowl Winner||Sugar Bowl Winner|
|Confidence - Watchability:||0.0807|
|Bowl||Date||Projected Winner||Confidence - Watchability|
|New Mexico||12/20/2014||Utah St||0.7864|
|Advocare V100 Texas||12/29/2014||Arkansas||0.7151|
|Popeyes Bahamas||12/24/2014||W. Kentucky||0.4061|
|Heart of Dallas||12/26/2014||Louisiana Tech||0.4038|
|Idaho Potato||12/20/2014||Air Force||0.2092|
|Quick Lane||12/26/2014||N. Carolina||0.2037|
Good luck in your bowl pools and happy holidays.
I dumped Massey Power into a bracket generator, added the autobids, tweaked it the minimum amount necessary to split teams in the same conference, and here's what came out:
There are probably some surprises here to anyone that hasn't been following any of the computer rankings: Iowa and Baylor in, Stanford in over Cal, Miami barely a 4 seed.
Michigan's a 4-seed, in a pod with Davidson, Marquette and the winner of a Colorado/Oklahoma play-in game.
The projected last 16:
- Indiana/St. Louis
- Kansas/Michigan St.
- Ohio St./Pittsburgh
- Georgetown/New Mexico
That top region is something else, if you think the ACC is a good conference this year. The computers don't--Sagarin has the ACC a distant fourth, barely better than the MWC.
Of course things might change a bit this afternoon, but (1) none of the computer sites will update their rankings until tomorrow and (2) my sense has been that the committee tends to ignore the late Sunday games because it's too late in the day to mess around with the bracket unless you're exchanging teams from the same conference. Sometimes they assign the two spots to the two teams and let the championship game decide which one goes where, but I'm pretty sure OSU > Wisconsin and Miami>NC and Florida>Mississippi regardless of what happens today.
Last week certainly simplified things. The leaders are done and in the legends we're left with:
- Iowa 84%
- Ohio State 38%
- Minnesota 91%
- Iowa 83%
- 2-0 32%
- 1-1 58%
- 0-2 10%
- 2-0 76%
- 1-1 22%
- 0-2 2%
Probability of winning the division: Nebraska 91%, Michigan 9%. That may sound grim but I'll bet it's a lot better than Mathlete's odds of winning on Saturday after NW got that last first down.
For what it's worth, Nebraska would be a one-point favorite over Wisconsin and Wisconsin would be favored by two over Michigan. At Sagarin Nebraska would be favored by half a point and M/W would be a toss-up.
Time to run these again.
Needless to say, this past weekend wasn't kind to Michigan on this front, as Nebraska dodged a bullet and except for Penn State the teams they'll face the rest of the way posted poor results.
The usual caveats apply--this assumes the probabilites are independent for each game, and while that becomes a better approximation as each week goes by it's completely ignoring (among other things) the possibility of a significant injury to a key player. That increases the likelihood of an extreme result and means the division-winning chances for the less likely teams are a little better than they appear here.
Massey win probabilities for the remainder of the season:
- Northwestern 64%
- Iowa 80%
- Ohio State 36%
- Penn State 67%
- Minnesota 92%
- Iowa 77%
- Michigan 36%
- Michigan State 44%
- Illinois 95%
That works out to:
- 3-0 18%
- 2-1 48%
- 1-2 29%
- 0-3 5%
- 3-0 47%
- 2-1 42%
- 1-2 10%
- 0-3 1%
- 3-0 15%
- 2-1 47%
- 1-2 36%
- 0-3 2%
Chances of winning the division are now 85% Nebraska, 13% Michigan, 2% Northwestern.
There are some bizarre possibilities still out there, like a 5-way tie at 4-4 (either Minnesota or MSU has to lose at least 5, since they still have to play each other). None of those possibilities let Minnesota or MSU win the division; they would always lose the divisional-record tiebreaker.
On the leaders side it's:
- Indiana 70%
- Ohio State 46%
- Penn State 51%
- Wisconsin 30%
- Penn State 27%
- Purdue 60%
Odds of winning the division are 77% Wisconsin, 23% Indiana.
Everyone else has been eliminated. There's still a possibility of a 3-way tie if Wisconsin loses out, Indiana loses both non-Wisconsin games, and Purdue or Illinois wins out. If it's Illinois they'll have lost to the other two and Indiana's win over Wisconsin gives them the tiebreaker and a trip to the championship game as a 5-7 team. If it's Purdue, the three teams will have split with each other, Indiana will be ruled out because they'll have the worst record in the division, and Wisconsin will then win the tiebreaker thanks to their win over Purdue.
I posted this on one of the weekend threads but the Massey site has now been updated with Saturday's results so here's a new version. Massey win probabilities for the remainder of the season:
- Minnesota 78%
- Northwestern 68%
- Iowa 81%
- Ohio State 37%
- Michigan State 51%
- Penn State 72%
- Minnesota 89%
- Iowa 71%
That works out to:
- 4-0 16%
- 3-1 43%
- 2-2 31%
- 1-3 9%
- 0-4 1%
- 4-0 23%
- 3-1 44%
- 2-2 27%
- 1-3 6%
- 0-4 0%
Chances of winning the division are now 71% Nebraska, 29% Michigan, ignoring the unlikely possiblity of someone else stepping in (I'm not going to run the numbers but eyeballing it I'd say Northwestern's chances are about 2-3%).
This may have been discussed here already, but what is the reaction to Massey short-arming that pass to the end zone? I want to check with the pious to see if we are filing this under “these are 21 yr old college kids, not professionals, they give 100% for UM and we should love them all no matter what they do on the field,” or if I'm a bad person because I got pissed.
Turns out I’m not even pissed though, its worse then that.
Slaughterhouse-Five, as many of you prob know, is a book about a guy named Billy Pilgrim, a WWII American POW in Germany who witnessed and survived the fired bombing of Dresden, along with other WWII POW-type carnage and atrocities. After Dresden the surviving Americans wander off out of the city but come back the next day to look for spoils, riding in a carriage. Upon arriving back in Dresden, it is pointed out to Billy by a couple of Kraut doctors that the horses pulling their carriage have neither proper bits nor horse shoes. Billy gets down and sees the bleeding mouths and cracked hoofs, and bursts into tears. It’s was the only time he cried during the war.
In my time following UM football, I’ve been wrapped in a fairly comfy cocoon of predictable emotion. Elation following important wins and blowouts. Annoyance following poor performances in a win. Applying a balm of perspective and all-time scoreboard to the stinging frustration of a loss, with varied results.
This season has exposed the M football fan to a series of numbing events and shocking imagery for which we simply are not equipped to deal. I watched UM have a FG blocked them muff a punt within 3 minutes of each other. Witnessed Zoltan get blocked. Saw NW 3rd and goal from the 17 converted. A white Northwestern receiver broke 10 yards open behind our coverage on a 53 yard touchdown pass. As I have since Toledo, I shuffled these events into the context of the game, loosely calculated their effect, and waited.
Then that goddamn pass to Massey. I’m not trying to go Lou Holtz or Nick Saban on you, nor am I trying to kick dirt in the face of a 21 year old kid. But when that pass went up I said, “touchdown.” Watching a senior’s hands go up for the ball (on senior day), then jerk back down as he tensed up, and north-freakin’-western intercepted the tying score... that was the play that finally got to me.