S&P+ 2016 win probabilities by game

Submitted by SAMgO on

As Bill Connelly continues to power through his Big 10 previews on SB Nation, we can slowly but surely fill out Michigan's projected win probabilities by game, and in turn our total wins projection. The only game that hasn't been released so far is OSU, which based on comparable opponents will probably be close to a coin flip. By using the very complicated Michigan=1-X, here's what the schedule looks like:

Opponent Win Probability
Hawaii 98%
UCF 96%
Colorado 93%
Penn State 75%
Wisconsin 80%
@ Rutgers 86%
Illinois 90%
@ MSU 55%
Maryland 87%
@ Iowa 67%
Indiana 86%
@ OSU              45-55%

If you give the OSU game 0.5 expected wins, that gets us to an expected win total of 9.63 on the year. Sweep the first seven games leading up to MSU and that number bumps up to 10.45, all else equal.

mickblue

July 25th, 2016 at 11:49 PM ^

I just have a very strong hunch we'll take them in E.L. this year. I had the same feeling going into Happy Valley in 1997. The win will be fueled by our defense, which has a chance to be special this year.

BursleyBaitsBus

July 25th, 2016 at 11:53 PM ^

Just gotta beat MSU and OSU at their respective places, never experienced that feeling as a student before. 

Hopefully, this year. 

USAFA007

July 25th, 2016 at 11:54 PM ^

Opponent Win Probability
Hawaii 100%
UCF 100%
Colorado 100%
Penn State 100%
Wisconsin 100%
@ Rutgers 1000%
Illinois 100%
@ MSU 100%
Maryland 100%
@ Iowa 100%
Indiana 100%
@ OSU 100%

Fixed it for you.

Gooseggs

July 25th, 2016 at 11:54 PM ^

There is such a difference in probability between @MSU and @Iowa. MSU lost more and the two teams were comparable last year. I would put the win probability at least equivalent between the two.

Ghost of Fritz…

July 26th, 2016 at 12:18 AM ^

or do others find it weird that the win probability for @Rutgers is just 11% higher than for the Penn State game and 6% higher than for the Wisconsin game?  I mean Rutgers is very bad. 

Kwitch22

July 26th, 2016 at 9:19 AM ^

Remember they are the birthplace of college football, so at one time they were at least in the top two programs in the country, so maybe back to there?

Ecky Pting

July 26th, 2016 at 2:01 PM ^

Win probabilities are merely an analog to the more traditionally applied point spread. As was referenced in the Big Ten 2016 Preseason Total Win Probabilities diary posted last week in regard to point spreads:

"The standard deviation ... is 15.5 points. That means that if a game shows a spread for your favorite team of -3 points, the outcome could be with high confidence within plus or minus 2 standard deviations of the point estimate, which is -3 ± 31 points in this case. So your favorite team could win by 34 points, or lose by 28!"

That, dear Poster, is why they play the games.

To put it in other words, if you bet on 1000 games between any two evenly matched teams (i.e. equal 50% win probabilities), 320 games would have more than a two touchdown difference; and 60 games, more than a four touchdown difference.

LSAClassOf2000

July 26th, 2016 at 6:42 AM ^

For some contrast, here's how Massey estimates the schedule as of, well, the preseason (their formula weights recent performance too, as I recall):

Hawaii - 99%

UCF - 99%

Colorado - 93%

Penn State - 80%

Wisconsin - 66%

at Rutgers - 85%

Illinois - 88%

at Michigan State - 48%

Maryland - 87%

ay Iowa - 59%

Indiana - 83%

at Ohio State - 28%

jdon

July 26th, 2016 at 10:09 AM ^

yesterday on Facebook the BTN used the Spartan punt block for victory to count down 38 days till games start...

to which I responded:Years from now they will be playing this clip repeating "the last time MSU beat michigan was..." it will take time for the Sparties to admit it but that was the apex. So I say to you, my spartan friends, enjoy and watch the clip ad nauseam because there will fewer and fewer highlights as the years role on. And michigan will not lose to michigan state in the next five years.