Weekly Game Probabilities Are Back

Game probabilities for week 4 are up at the New York Times. With the demise the Fifth Down Blog, the probabilities are now on the NYT website proper. This week I explain why the randomness of turnovers makes the Giants-Chiefs game a closer match-up than their win-loss records suggest.

In retrospect, turnovers explain a great deal of a team’s fortunes. But prospectively, team turnover statistics don’t predict game outcomes as much as you might think. The reason behind this distinction is something called auto-correlation. Put simply, turnovers are very random. Only a small portion of a team’s past turnover rate carries forward to be predictive.

8 Responses to “Weekly Game Probabilities Are Back”

1. MFLoGrasso says:

I'm curious about how the Bills are the third-largest favorite this week, with a 0.70 win probability. By your own EPA/G stats, the teams are about even defensively, and Baltimore is better offensively. Vegas has them as a home dog this week, and while I would love to think there is little to worry about at One Bills Drive this Sunday, I have trouble fathoming them having such strong odds of winning.

2. Nate says:

> ... I would love to think there is little to worry about at One Bills Drive
> this Sunday, I have trouble fathoming them having such strong odds of
> winning.

I think the graphic at the top of the page is retrospective, in his predictive rankings (http://www.advancednflstats.com/2013/09/team-efficiency-rankings-week-3.html), Buffalo is significantly stronger than the Baltimore. I'm sure that home field advantage also comes into play.

That said, NFL football is very random, and there's only been a small sample of games.

3. Anonymous says:

Is there a reason you've switched from the old text-based tables to an image for the probabilities table? Totally invisible to a screen reader.

4. J.R. says:

The Bills' ticket to victory is to get deep in the hole early so they have to start throwing deep all the time; eventually those high-variance plays will beat our mediocre safeties. They're not going to run it against us. If the Texans' short passing game can't score on us then neither can the Bills'.

Conversely, what do the Bills have on defense that should scare the Ravens? We beat the Browns and Texans, and put up 27 against the Broncos despite losing.

5. Anonymous says:

yeah, the Bills game seems way off, as does the Lions being strongly favored over the Bears. It may be just another lesson in small sample sizes ... we'll see.

6. Brian Burke says:

I don't doubt that .70 is too high for BUF. The thing is...just look at BAL's numbers. If they're going to keep trying to run, they better start doing a better job.

The new graphic version of the table is unfortunate, but it's required now that we're off Fifth Down and on the full NYT site.

7. Brian Burke says:

I don't doubt that .70 is too high for BUF. The thing is...just look at BAL's numbers. If they're going to keep trying to run, they better start doing a better job.

The new graphic version of the table is unfortunate, but it's required now that we're off Fifth Down and on the full NYT site.

8. Anonymous says:

You were right about BUF, CLE, and DET. Kinda amazing.