Why Sportswriters Didn't Become Statisticians

by rrusczyk, Nov 21, 2009, 6:40 PM

Exhibit Number 1.

See if you can pick out all the dodgy uses of stats, and all the outright mathematical inaccuracies.

(That said, all the "Belichick is a genius" and "Belichick is a moron" pieces seem silly to me -- I don't think either possible call clearly dominated the other. It did at least overshadow del Rio's call, which was just as interesting...)

football

statistics

Comment

4 Comments

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

Haha I read that a few days ago. Or it may have been yesterday. I can't remember anything much these days...I blame school.

by xpmath, Nov 21, 2009, 9:50 PM

Report

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

Unfortunate. At least sabermaticians (SABR magicians?) know what they're doing, for the most part.

by CA Math, Nov 25, 2009, 5:59 AM

Report

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

Read this response: http://sports.espn.go.com/espn/commentary/news/story?page=macgregor/091123

by SamE, Nov 27, 2009, 4:40 AM

Report

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

My fantasy football league consists of 3 Ph.Ds (probability, statistics, algebraic geometry), 3 Ph.D students (all probability), and an M.A. in math (not sure which field). Needless to say, this column was the topic of a lot of discussion on the league message board this week.

It may be the single worst article I have ever read from a mathematical viewpoint. It wasn't exactly news to me that Bill Simmons is an idiot (I believe that one of our league members emailed Simmons from his University of Chicago email to tell him exactly how much of an idiot he is), but he's outdone himself here. The two best (worst) parts of the article are the following:

Bill Simmons wrote:

I am not disputing the numbers or the methods for achieving them. But by Monday night, based on various columns and message boards (as well as e-mails to my reader mailbox), you would have thought Belichick was a genius for blowing the game. He played the percentages! It wasn't as crazy as it looked! By this logic, Belichick also should have held a loaded pistol to his head on the sideline, spun the chamber and tried to shoot himself like Chris Walken in "The Deer Hunter." If those 1-in-6 odds came through and he succeeded, we could have said, "Hey, he played the percentages: 83.6666 percent of the time, you don't die in that situation! You can't blame him for what happened!"

It should be pretty clear that this calculation is only reasonable if the loss of utility you associate with dying is 0.

Bill Simmons wrote:

According to a formula called "Expected Win Probability When Going For It,"

This is also known as 'a version of Bayes' theorem so trivial that a retarded monkey could come up with it'

Anyways, lost in all of Bill Simmons' confusion about expected values, he (implicitly) makes a good point: the ESPN 'win probability' referred to in the article is absolutely useless, and estimating the probability of winning via Bayes' theorem is only a meaningful exercise if you can estimate the probabilities of certain events well. In particular, lousy hypotheses (like estimating the probability of every event by the league averages) lead to lousy conclusions.

Unfortunately for Bill Simmons (and, sadly, a lot of other {sports personalities, politicians, influential public figures}), probabilities don't really exist in any meaningful sense. They are all 0 or 1. And that means that you can always be results-oriented and look back on what actually happened, and conclude that since it happened, the probability of that event must have been 1 in the first place.

by blahblahblah, Nov 27, 2009, 9:20 AM