Writers of Pro Football Prospectus 2008

24 Nov 2008

2008 Quick Reads: Week 12

All yards are not created equal, and Brett Favre's presence at the top of the Week 12 rankings are a sign why.

Posted by: Bill Barnwell on 24 Nov 2008

25 comments, Last at 25 Nov 2008, 5:16pm by No Biases


by Carlos (not verified) (not verified) :: Tue, 11/25/2008 - 12:05am

Your pro-Pats, anti-Favre biases are getting really old.

by Bowl Game Anomaly :: Tue, 11/25/2008 - 12:14am

You forgot to mention Aaron's jealousy toward the GB secondary (which is being absolutely torched by Drew Brees as I type).

(Formerly "The McNabb Bowl Game Anomaly")

by Bobman :: Tue, 11/25/2008 - 1:21am

Agreed. Disgusting. They should be ashamed of themselves.

by dbostedo :: Tue, 11/25/2008 - 11:11am

Wait - is that a joke? I don't get it... I'm looking for the bias, and I can't seem to find it...examples?

by Brian (not verified) :: Tue, 11/25/2008 - 12:56am

The article trumpets about how your stats account for game situation and how each yard gained affects a team's prospects of winning. Have you considered using win probability added (WPA) instead of your current system of stats? Wouldn't that be a simpler, more accurate way of measuring what you're interested in? You could adjust it for opponent strength too.

by Rich Conley (not verified) :: Tue, 11/25/2008 - 10:58am

No, WPA wouldn't work better. A 15 yard pass in the 4th quarter down 24 does still say something about a team. In WPA, that pass would just go away. We're already dealing with small sample sizes. Why toss out plays?

by MJK :: Tue, 11/25/2008 - 12:31pm

I'm not entirely clear on how WPA would be calculated (you'd look at the probability of winning before and after each play and how it changed?), but given that plays in the first quarter do very little to affect win probability, and plays in the last quarter of a close game have a huge effect, wouldn't you end up over-emphasizing 4th quarter plays in close games? For example, imagine you're out of timeouts, have the ball and about 30 seconds left, down by 4, facing 4th and 10 on the other team's 15. You probably have about a 10-15% chance of winning there (figure a 25% chance of converting the 4th down, and maybe a 40-60% chance of punching it in after that). If you fail to convert the 4th down, your win probability just went from 10-15% to zero, meaning that the opposing defense gets 10-15 points of WPA for forcing one failed play. But if the team is in the same situation at the half, you get maybe 1 point of WPA, since being down by 4 is not all that different than being down by 1 or up by 3 at the half. In the grand scheme of things, those seven points are just as important to a close game no matter when they are scored.

by Yaguar :: Tue, 11/25/2008 - 1:10am

I'd like to know the DYAR numbers for Drew Brees from Monday Night.

by Bobman :: Tue, 11/25/2008 - 1:20am

I wouldn't. Hurt my head.

I am surprised Peyton Manning is so low--doesn't his crazy 3rd down conversion rate boost his rating way up? I guess the SD D ranking drags him backwards.... The opposite of Favre's Titan effect.

I also would have thought Anthony Gonzalez would have the same things for and against him--a TD, something like 5 of 6 on 3rd down receptions for conversions, but... that SD D again. Not sure what the FO software sees, but they looked plenty good to me last night.

by STI (not verified) :: Tue, 11/25/2008 - 2:07am

"I am surprised Peyton Manning is so low--doesn't his crazy 3rd down conversion rate boost his rating way up? I guess the SD D ranking drags him backwards.... The opposite of Favre's Titan effect."

Yes, the strength (or lack thereof) of the D plays a role, but I think his "crazy 3rd down conversion rate" is balanced out by the fact that the Colts ended up with 17 third-down plays out of 68 (25%) to begin with.

By contrast, Matt Cassel and the Patriots only had 8 third downs out of 70 (11%). [Both teams had two fourth-down plays, and converted both.]

by zlionsfan :: Tue, 11/25/2008 - 10:09am

"couldn't" care less.

Please, think of the children.

So the Lions manage to contain Garcia somewhat, limit Cadillac to fewer DYAR than any other RB last week ... and still get crushed.

You know, I have a much greater admiration now for people who admit that they're overmatched and move on, rather than clinging stubbornly to a position that is well outside their range of competency for more than seven years.

by tuluse :: Tue, 11/25/2008 - 12:28pm

You know it took me a second to realize you were talking about Millen. At first I thought, you mean the Lions should petition to leave the NFL?

by Daniel (not verified) :: Tue, 11/25/2008 - 10:45am

Where would Drew Brees' game last night rank? He was out of control.

by Rich Conley (not verified) :: Tue, 11/25/2008 - 11:00am

"Manning's day could have been even better had his receivers not dropped a couple of those seven incompletions"

At the same time, his day could have been a lot worse if his receivers hadn't caught a handful of balls thrown 4 feet over their heads.

by Ari (not verified) :: Tue, 11/25/2008 - 11:07am

Funny how the article laments how baseball statistics and hitting quality measures don't adjust for the thin air in Colorado... yet, as I recall, DVOA doesn't adjust for weather either.

Did I miss something here?

by Tom Gower :: Tue, 11/25/2008 - 12:16pm

Yes-Aaron writing that he's tried to adjust the general DVOA formula for weather, and that doing so has not provided useful information. Weather (and altitude) is incorporated into special teams DVOA, since it clearly does matter there.

by tally :: Tue, 11/25/2008 - 12:18pm

Special teams DVOA does in fact adjust for stadium, mainly for indoor/outdoor and altitude. FG kicking and kickoffs are obvious beneficiaries of kicking in Denver, for example.

I don't know if it has been conclusively shown that weather has a significant effect on other areas, although intuitively, I would think so. Perhaps someone can point me to a study by FO?

by tuluse :: Tue, 11/25/2008 - 12:32pm

I think their was an article where they showed that cold didn't have an consistent effect, but wind speeds had a large effect. Unfortunately wind speed information is difficult to get a hold of.

by MJK :: Tue, 11/25/2008 - 12:39pm

Yes, I remember that too. Neither cold, nor rain, nor snow, nor sloppy fields affected much, but wind speeds in general degraded passing stats across the board. Not exactly surprising, but not particularly useful, since wind conditions are not well reported.

I remember we were speculating that the reason why field conditions other than wind didn't seem to affect things was because it was hard to determine if a sloppy field would affect the offense or the defense more...it's harder for WR's to run routes and RB's to get their footing, but it's harder for DB's and LB's to make cuts while covering guys or plugging cutback lanes, and the pass rush is slowed.

by Tundrapat (not verified) :: Tue, 11/25/2008 - 1:57pm

It doesn't appear to come up that often, but I'd say inclement weather plays a part in the entire game, not just the special teams.

Recall the Pittsburgh-Cleveland game in Week 3 this year. Rain, strong wind gusts, and both offenses look horrid (it appears to affect the offense more than defense, since pass routes are largely based on a receiver being in a location at a precise moment, and the QB getting the ball to the exact same spot). Final score: 10-6. Steelers had 281 total yards, Browns had 208 total yards.

Now recall last year, Miami at Pittsburgh. Torrential rain, that hilarious punt attempt where the ball's forward momentum simply stopped in midair, and hit the ground without bouncing once. Final score, 0-3. Steelers had 215 total yards, Dolphins had 159.

So based on a small sample size, it appears that: a) weather can affect the entire game, and; b) weather appears to affect the offense more than the defense.

Then again, maybe it just affects the Steelers' games.

I suppose one could look at the effects of the 'Snow Bowl' and other rain- and snow-laden games. Strangely, snow doesn't seem to have as much effect on games as heavy rain (based entirely on my recollections of snow games).

by tally :: Tue, 11/25/2008 - 12:14pm

Maybe we should start taking wagers on what Brees' DYAR is from last night. I say it's higher than Favre's and maybe by a decent margin. GB's pass D is nearly as highly rated as the Titans' D, Brees' completion % is about the same but with a higher YPA. The Saints didn't face too many 3rd and 4th downs because they were so good on 1st and 2nd. He threw no INTs and only took one sack.

by Joseph :: Tue, 11/25/2008 - 3:11pm

(Saints homer)
I wondered the same thing. Of course, two 70 yd. TD's didn't hurt--I'm sure that all Brees fantasy owners were happy too (or at least those NOT residing in Wisconsin). Also, those who were decided to play Lance Moore were happy too (25 points!! in standard scoring).

I bet that Brees would be #1, Moore would crack the top 5 WR, and Pierre Thomas would crack top 10 RB (but prob. not have enough DYAR for top 5 because GB doesn't have a good rushing D, and he only had 15 attempts IIRC).

by BucKai :: Tue, 11/25/2008 - 12:40pm

I cannot see how one can write "context was left out of the discussion -- despite being crucial to establishing the validity of the measure." and "Jeff Garcia....What those numbers don't show is that Garcia was sacked six times and fumbled once."

If context is the goal, then the fumble and one of the sacks belong to Cadillac Williams. He was rusty, in the wrong place and hit his own QB. Not to focus on the negative, give him credit for putting six points on the board. Heck, he was Detroit's most productive RB.

by Tundrapat (not verified) :: Tue, 11/25/2008 - 2:00pm

Productive on the field, sure...but did Caddy take anyone else's bags when he left Detroit?

by No Biases (not verified) :: Tue, 11/25/2008 - 5:16pm

Your calls of bias are ridiculous. Barnwell is a Giants fan.