Weekly Probabilities - 11/11

Goalscorer9's picture
November 11, 2013 at 1:56p
19 Comments

Here they are!  Again, although the process is objective, I acknowledge my probabilities of individual games are subjective, though I think this gives us a pretty good picture of where we are at.

As always, my numbers will be attached for anyone who would like to critique.

Assumptions:
-Ohio State finishes undefeated.  This whole argument starts with "If Ohio State goes undefeated", so we'll assume they do to calculate probabilities
-If Alabama or Florida State finish undefeated, they will remain ahead of Ohio State.
-No 1 loss team will finish in front of an undefeated Ohio State
-Baylor, Fresno State, and Northern Illinois will not jump Ohio State, even if they finish undefeated

Numbers you care about:

-Probability Alabama wins out: 44.1%
-Probability Florida State wins out: 52.4%

-Probability all of them lose, and Ohio State finishes numero uno: 26.7%

-And most importantly, the probability that at least 1 of these teams lose, and an undefeated Ohio State team finishes in the top 2: 76.9%

76.9% is a number I'll take at this point.  Probably won't change much (for good or for bad) until November 30th when Bama takes on Auburn and FSU plays Florida.  We just need one more slip up, and the numbers tell us that is fairly likely.

For those of you who would like to include the possibility of Ohio State losing:
-Probability of Ohio State winning out- 50.1%
-No undefeated assumption odds of Ohio State being in the national championship game- 38.5%

 

Comments

gobucks96's picture

Think you can increase the odds that FSU beats Florida up a bit.
 
Well done again.

Goalscorer9's picture

Haha, perhaps irrationally, I just think a rivalry game AT the swamp, I can't put it up above .75.  Also am sort of hoping the gators have healed up a bit by then, but I don't know if that will be the case

d5k's picture

Right, if we are 80% to beat Michigan, then FSU is probably 90% to beat Florida.  They will probably be a 20+ point favorite.  Florida isn't going to make a bowl game.

Goalscorer9's picture

yeah I'll probably put it in the 80's.  I think 90 is a little too high.  I'll see how the teams look this weekend

Goalscorer9's picture

Eh I actually just went to a gator site (something I typically try to avoid) and it looks like most of their injuries are season ending.  Horrible luck for them this year.  I almost feel kinda bad.

That game may get a little bump in FSU's favor when I re assess next week.

Killer nuts's picture

I enjoy these write ups by you and certainly there is some arguments to be made regarding each probability that you use. A couple that standout for me:
-I'm not sure that Alabama is more likely to lose the SEC title game on a neutral field to Mizzou/S Carolina than they are to lose on the road at #7 Auburn in a rivalry game
-I also am not sure why you're more confident that we will win at 6-3 Michigan than you are that Florida State will win at 4-5 Florida.
Small adjustments in the probabilities won't change your numbers a great deal but those 2 stuck out to me. Thanks again for doing these, they help put in perspective the difficulty of going undefeated. It might also be interesting to look back to the beginning of the season for these 3 teams, my guess is that the likelihood of these 3 all being undefeated at this point in the season is less than 5% and yet here we are

d5k's picture

The likelihood of any scenario is tiny if you look back far enough.  

Goalscorer9's picture

Killer,

Thanks for the feedback. To address the thought process:

1) I just think Auburn doesn't match up very well with Alabama.  I've only watched Auburn a couple times this year, but they seem very 1 dimensional on offense, and that dimension is the run game, which I think plays into Bama's hands.  That's why I have them as more likely to lose in the SEC championship game.  It's really hard to compare right now since we don't even know who their opponent would be.  I'm doing the best I can, but it's kind of a shot in the dark with some of these haha.

2) to put it simply, I think Michigan is very lucky to be 6-3, as they should probably be 4-5 with losses to Akron and UCONN.  Michigan looks really, really, bad right now.  As discussed above, I probably need to up the percentage for FSU vs Florida, and will do so next week.  I wasn't aware just how bad the injury situation was for the gators.

I tweak the individual numbers each week, and I will definitely take everyone's feedback into account when doing so!

d5k's picture

I know it is just an assumption but while I don't think a 1-loss team has a reasonable shot of passing us and getting into the NCG, Baylor has a realistic chance to pass us.  I think the computers will narrow as they are already starting to and it will be up to the hype machine to decide who to promote or who has the most style points after Baylor actually plays the rest of the good teams on its schedule in November.  They made up a ton of ground in the computers and it looks like OSU has about a 60/40 advantage in the polls which isn't insurmountable at all.  A bunch of voters probably just watched Baylor for the first time all year.

Goalscorer9's picture

I kind of touched on Baylor in last Friday's post.  If it starts to look like they're going to pass us, I'll include them in ensuing posts.  For now, it's impossible for me to calculate the odds of whether or not an undefeated Baylor finishes in front of an undefeated Ohio State, so I'm going to ignore it for now, as we still look like we're in good shape.  I don't *think* we have to worry about it, but we will see.

d5k's picture

It really is the odds that they will look better than us to enough of the pollsters.  I think the computers will narrow as the quality of their average opponent approaches ours as their quality opponents in November start to cancel out all the cupcakes they have played.

Goalscorer9's picture

yeah but I can't place odds on human opinion lol

Johnny_Buck71's picture

This is why I love 11W and this kind of post in particular. There is logic imposed from the start. The critique is sound and impersonal. The original poster is responsive to suggestions and puts forth the effort to further research various aspects not originally considered. In other words, an intelligent post that attracts an intelligent fan. Good job!

A nickel ain't worth a dime anymore. - Yogi Berra

branz's picture

76.9% seemed a little high to me, so I did some calculating on my own. Based on you game by game probabilities, I agree with your assessment of Alabama and Florida State's individual probabilities of going undefeated (44.1% and 52.4%, respectively). I don't agree with the probability of the following two scenarios:
Alabama AND FSU both lose, OSU undefeated (I get 13.3%)
Alabama OR FSU loses, OSU undefeated (I get 38.5%)
I believe you forgot to factor in Ohio State's chances of going undefeated in you equations. Let me define my variables first...
     A = Alabama's chances of going undefeated = 44.1%
     F = FSU's chances of going undefeated = 52.4%
     O = OSU's chances of going undefeated = 50.1%
Here is how I calculated each...
     Alabama AND FSU both lose, OSU undefeated
     = (1 - A) x (1 - F) x O
     Alabama OR FSU does not go undefeated, OSU undefeated
     = (1 - A x F) x O

Johnny_Buck71's picture

Regarding probabilities...
Isn't it [P(A) losing OR P(F) losing] AND P(O) winning out?
where P(A), P(F) and P(O) = ALA, FSU, and OSU winning out, respectively.
therefore, probability either FSU or ALA lose and OSU wins = [(1-A) + (1-F)] x P(O)

A nickel ain't worth a dime anymore. - Yogi Berra

branz's picture

Only if the events are mutually exclusive, which they are not. In that case, the probability of both teams losing would need to be subtracted (removing the intersection of both teams losing).
I chose to calculate P(A undefeated) AND P(F undefeated). Subtracting that value from one gives the probability of one of the two losing. 

Goalscorer9's picture

In the post, I say that my calculations are based on the assumption OSU goes undefeated.  People want to know "if we got undefeated, what are our odds?" this is what I am reporting.

branz's picture

Thanks for pointing that out. I'd give you a couple upvotes if I could.

Goalscorer9's picture

Note, in the last line I showed the "no undefeated assumption number" and we get the same value.