Massey predicts BYU will go 10-2

[jay santos]

I was going to make a post like this when people were talking about predictions for total W-L this season. People generally look at the schedule and say something like, "we have 8 easy games, four tough games, we'll split those, we'll go 10-2". When those 8 "easy" games our probability might be 80%, which means an expected 6.4 wins out of those 8 games.

Another way to look at this, is that when Boise State goes 12-0 against easy competition, people mistakenly say, "well the best team they beat is #30, so we really don't know if they're any better than the #25 team". When you go 12-0 like that even against teams that bad, it means that most likely your true power rating is very, very high. Because even a top 15 team would likely lose a game with that schedule.

[Indy Coug]

Cool. Simple Monte Carlo simulation I presume?

You could also assume independence between games and just sum the probabilities to get an expectation of 8.22 wins and 3.78 losses.

[Indy Coug]

Another way to look at this, is that when Boise State goes 12-0 against easy competition, people mistakenly say, "well the best team they beat is #30, so we really don't know if they're any better than the #25 team". When you go 12-0 like that even against teams that bad, it means that most likely your true power rating is very, very high. Because even a top 15 team would likely lose a game with that schedule.

Exactly

[YOhio]

People, be careful out there when it comes to probabilities.

Seriously, guys. Be careful. Bayes theorem. Learn it.

[Indy Coug]

Seriously, guys. Be careful. Bayes theorem. Learn it.

Embrace it

[BigPiney]

This was very enlightening. Thanks. Makes that graduate level stats class I took years ago make some sense.

[pelagius]

You could also assume independence between games and just sum the probabilities to get an expectation of 8.22 wins and 3.78 losses.

Of course, I wanted to do the distribution. But I did forget to mention there is an easy way to get expected wins.

[pelagius]

Massey's newest rankings give some different numbers than those in the first post.

http://masseyratings.com/cf/compare.htm

In these rankings, BYU is 27, and:

Weber is unlisted
Utah is 30
Boise is 22
Hawaii is 97
USU is 68
Oregon State is 69
ND is 18
Georgia Tech is 41
Idaho State is unlisted
SJSU is 80
NMSU is 110.

I don't know how to figure the probabilities from that.

Those are masseys composite average of a whole bunch of models. So you're really referring to a different model.

[All-American]

Those are masseys composite average of a whole bunch of models. So you're really referring to a different model.

Ah.

[Rickomatic]

I won't make much credence in the probabilities of a prediction model that can't even get all of the opponents correct. BYU plays Idaho in November. They played Idaho State last year. But, with that being said, Idaho State maybe the better of the 2 teams this year.

[Jeff Lebowski]

I won't make much credence in the probabilities of a prediction model that can't even get all of the opponents correct. BYU plays Idaho in November. They played Idaho State last year. But, with that being said, Idaho State maybe the better of the 2 teams this year.

That was AA, not Massey. See link in Pelagius post.

[statman]

I just threw together a quick Monte Carlo analysis with 5000 instances, just for fun. Not quite to statistical convergence, but close. This is what I got. Looks like your equations were correct.

I am not liking our chances for an undefeated season.

Every thousand years, two undefeated seasons. 1984 & 2012. No problems there!

[Indy Coug]

I just threw together a quick Monte Carlo analysis with 5000 instances, just for fun. Not quite to statistical convergence, but close. This is what I got. Looks like your equations were correct.

I am not liking our chances for an undefeated season.

Even if you have a 95% probability of winning each and every game, you're a coin flip on whether or not you go undefeated.

[pelagius]

I won't make much credence in the probabilities of a prediction model that can't even get all of the opponents correct. BYU plays Idaho in November. They played Idaho State last year. But, with that being said, Idaho State maybe the better of the 2 teams this year.

That was AA, not Massey. See link in Pelagius post.

Leb's right, of course, but let me add a little more about the point of the thread. My original post and Leb's follow-up are not meant as attempts to seriously get the probabilities right (only relative to the model's predicted probabilities) for different outcomes this season. The general point was illustrate what was well expressed by Jay earlier in the thread:

I was going to make a post like this when people were talking about predictions for total W-L this season. People generally look at the schedule and say something like, "we have 8 easy games, four tough games, we'll split those, we'll go 10-2". When those 8 "easy" games our probability might be 80%, which means an expected 6.4 wins out of those 8 games.

Another way to look at this, is that when Boise State goes 12-0 against easy competition, people mistakenly say, "well the best team they beat is #30, so we really don't know if they're any better than the #25 team". When you go 12-0 like that even against teams that bad, it means that most likely your true power rating is very, very high. Because even a top 15 team would likely lose a game with that schedule.

Second, let me add what I have stated earlier in this thread (it feels like at least 3-4 times) that no one should take a any computer model very seriously after one game (it is usually mostly some weighted average of end of year rankings from the previous 5 years or so with usually a lot of weight on last year). That said, a ranking of around 30 for BYU, is not unreasonable. For example, the current oddsmaker top30 ranking has BYU ranked 29th (which I would probably trust the most at this point). Personally, my 90% confidence interval for BYU is, 10-40.

[Tim]

I won't make much credence in the probabilities of a prediction model that can't even get all of the opponents correct. BYU plays Idaho in November. They played Idaho State last year. But, with that being said, Idaho State maybe the better of the 2 teams this year.

[YOUTUBE]iApz08Bh53w[/YOUTUBE]