Writers of Pro Football Prospectus 2008

05 Jan 2009

2008 Quick Reads: Wild Card Weekend

How did Darren Sproles get -2 DYAR? Well, he did the worst thing a back can do with the ball in his hands.

Posted by: Bill Barnwell on 05 Jan 2009

31 comments, Last at 08 Jan 2009, 12:11am by beargoggles


by Matt W (not verified) :: Mon, 01/05/2009 - 4:28pm

This brings up the question: When did Marcus Pollard sign with Atlanta?

THANK YOU for not saying "begs."

by Kevin from Philly :: Mon, 01/05/2009 - 4:58pm

You failed to add that the Philly WRs practically carried Westbrook from the 40 yard line on. Great job of blocking, especially the rookie DeSean Jackson. That said, B-West is still the man!

by KinSD (not verified) :: Mon, 01/05/2009 - 4:58pm

Is there a way of calculating how much DYAR he had for his punt and kickoff returns? Because there was tremendous value there. Thanks.

by Bobman :: Mon, 01/05/2009 - 5:15pm

Reggie Wayne had a couple significant drops early. Ouch.

Go, Edge, go!

by The Blow Leprechaun (not verified) :: Mon, 01/05/2009 - 7:12pm

The low DYAR figure for Sproles is interesting. Watching the game, I expected him to have put up something like 150-200 yards because he just seemed to be everywhere, and picking up 10 yards every time he touched the ball. I was shocked when I looked up his actual numbers from the game. I wonder why my perception was so different from reality.

by navin :: Mon, 01/05/2009 - 7:13pm

Did Antonio Gates' fumble get counted in his DYAR? I heard that the play by play missed it. If he loses that fumble the Colts probably win that game.

by randplaty (not verified) :: Mon, 01/05/2009 - 7:18pm

So are we saying that if Sproles didn't play on offense at all, the Chargers would have been better?

by Love is like a bottle of gin (not verified) :: Mon, 01/05/2009 - 9:23pm

You do not understand DYAR at all of you think this. It has little to nothing to do with individuals.

by Sergio :: Tue, 01/06/2009 - 12:16pm

And yet, you did *nothing* to correct his notion. Congratulations.

As I understand it, it just means that Sproles was less productive than a hypothetical average replacement from the bench ("replacement player"). 20-something yards less productive (or whatever his DYAR was).

In any event, I'm pretty happy to see Pennington something other than crucified for his performance. People really forget this was the best defense in football, and that the Dolphins didn't have any top-flight receiver talent to make the Ravens pay for consistently putting pressure on Pennington.

Not that 4 INTs are a good performance, mind you. It's just not "oh my god, fire him yesterday!" bad, considering the opposition.

-- Go Phins!

by Dales :: Mon, 01/05/2009 - 7:56pm

This brings up a question for me.

"Everyone knows that fumbles are bad. Longtime readers of Football Outsiders know that fumbling the ball, even if you end up recovering it, is a bad sign because fumble recovery isn't a skill -- that is to say, there's no team year-to-year which recovers a consistently high or low percentage of the balls that hit the ground."

And I believe it is well established that while recovering a fumble is not a repeatable skill, fumbling itself is one.

But is the propensity to fumble at certain portions of the field repeatable? In other words, do some people fumble more at the goal line than they do at mid-field?

The reason I ask is that DVOA, if I am recalling correctly, treats fumble recoveries as if they are random, rather than relying on the fact that being random over time they will equal out. But if *where* a fumbler fumbles is random, then does it make sense to hit someone with a huge penalty for fumbling on the one? Sure, over time it will likely even out, but does that make things more accurate from a 'true level of performance' perspective and making DVOA ratings more accurate?

by tylerdolphin :: Mon, 01/05/2009 - 8:27pm

With a fumble on the one, you are leaving sure points on the field pretty much. a fumble on the 50 isn't as sure of a thing to cost you points. And I think DVOA measures a players value to a team. A player costing his team sure points will be less valuable. Thats my impression of that anyway.

by Dales :: Mon, 01/05/2009 - 9:48pm

By that logic, a player who recovers a fumble on the one is giving his team points and is therefore valuable, but DVOA does not work that way because fumble recovery is random.

My point is if *where* a fumble happens when it happens is random, then penalizing a player more for it happening on the one might not make the system more accurate in regards to determining the true ability of teams.

by Pat (filler) (not verified) :: Tue, 01/06/2009 - 12:18am

The value of a turnover doesn't change that much over the length of the field. There's a small correction, but not a huge amount - if you fumble on your opponent's 1, you've given up a sure score, but you've also put the opponent in horrible field position. Fumbling on your own 1 is horrible (sure score for the opponent), but you were already in a horrible position to begin with.

There's an old FO article about this. Hidden Game of Football proposed that turnovers were all worth -4 points, regardless of position. If you look into it with a bit more detail, turnovers near the edges are worth a bit more, near the center of the field a bit less. But it's something like a 10% correction.

What down it is has a much larger effect on the value of a turnover than field position.

by andrew :: Tue, 01/06/2009 - 1:38am

Is causing fumbles a repeatable skill?

Are fumbles versus teams that cause a lot of fumbles less damaging than versus ones that don't?

by Pat (filler) (not verified) :: Tue, 01/06/2009 - 5:26am

Is causing fumbles a repeatable skill?

Mother of God, yes. Have you never seen Brian Dawkins play?

by Theo :: Tue, 01/06/2009 - 8:09am

No. (if it were different, I'd even say it's more damaging.)

by andrew :: Tue, 01/06/2009 - 9:18am

Alright, yes I know some players cause more fumbles.

What I meant was, in terms of DVOA, are they more damaging to one's rating?

e.g., having a poor day against an elite run defense doesn't result in as low a score as having one against a porous run defense.

So i was wondering if fumbling against, say, the Ravens or Eagles would not hurt a RB's score as much as fumbling against, say, the Lions (I'm assuming they cause less fumbles).

by hector :: Tue, 01/06/2009 - 6:05am

If you'd bet six weeks ago that the Cardinals would hand the ball off to James on the first three plays of their wild-card game, you could've won a lot of money.

I guess you joined the game a little late (and judging from the BS Report, Aaron did too). Check the game log, Arizona threw on its first three snaps of the game (and then punted). The Edge Experience was featured in Drive 2.

by hector :: Tue, 01/06/2009 - 6:16am

So according to these stats, Tomlinson was more valuable to the Chargers than Sproles Saturday? Epic. Fail.

by Theo :: Tue, 01/06/2009 - 8:23am

His yards were more valuable, because Sproles negated a lot of yards by fumbling the ball away. (that costs about 2 drives of yards.) and...
No you're right, it makes no sense. At all.
Sproles is a superb player who had a game that I will really remember as one of the best I've seen. Especially the way he did it. Grabbing every opportunity he was given. That mind set also lead to the 'stretch for a TD on second down' mistake. That is smart on 4th down, not so smart on 2nd.
I also don't understand the catching YAR. He was mostly the dump off receiver. So making yards is only a bonus, because it's because of him it isn't a sack/intentional grounding/incompletion/INT. He can't be blamed/punished for NOT making a first down.
All that doesn't take away that he was the MVP of the week.

by hector :: Tue, 01/06/2009 - 10:42am

It's a case of the Burden of Playing a Lot - you have more time to mess up and get graded down for it. Meanwhile, there's the D.J. Hackett problem - a part-timer plays well in a small sample and gets overly praised for it. Then people act surprised when the Hacketts of the world don't keep their success rate afloat as starters.

by Eddo :: Tue, 01/06/2009 - 1:27pm

It's DVOA, not DYAR, that overrates players like Hackett. Remember, DYAR is cumulative; Devery Henderson led the league in WR DVOA this year, but he was 17th in WR DYAR.

Look at Michael Turner. His RB DVOA is only 4.2%, good for 19th in the league, but he's 6th in RB DYAR. This is because he had a lot of touches. It works both ways.

And maybe Sproles getting worse with more carries is a sign that he's really not all that valuable. After all, the best RBs maintain high value despite having many carries.

by Jon Coit (not verified) :: Tue, 01/06/2009 - 12:30pm

Sproles really had a boom and bust day, with quite a few stuffs. I think folks are understandably entranced by his speed and ability to make other people miss...but he's not the kind of player that can turn a 2-yard loss into a 3-yard gain by lowering his shoulder. The times he attempted that Saturday he got crushed.

...ok, looked at the pbp. Sproles was credited with 23 carries, 2 were no gains with defensive penalties. 11 of the remaining 21 carries got 2 or fewer yards, 2 were for 3 yards. Then 8 carries for the remaining 88 yards; a few of those were precisely the kind of runs DVOA catches, e.g. an 8-yard run on 3rd and 10 which set up the game-tying field goal. Leaving out the fumble, it looks like not much better than league-average.

by hector :: Tue, 01/06/2009 - 12:50pm

he's not the kind of player that can turn a 2-yard loss into a 3-yard gain by lowering his shoulder.

Do you know a lot of backs who routinely pick up five extra yards with a shoulder drop and sheer power?

ok, looked at the pbp. Sproles was credited with 23 carries, 2 were no gains with defensive penalties. 11 of the remaining 21 carries got 2 or fewer yards, 2 were for 3 yards. Then 8 carries for the remaining 88 yards

A lot of successful running back days fit this profile, all around the league. Rushing 20 times for 100 yards is very seldom about going 4, 6, 6, 4, 4, 6, etc. Just about any runner who gets 20 or more carries will have a bunch of plays in there with modest or no gains.

I accept that the effectiveness of Sproles's night is probably less than is commonly held by the average fan. But I stand by the opinion that it's ludicrous to suggest that LaDainian Tomlinson was more valuable to the Chargers on Saturday night than Darren Sproles was.

by Jon Coit (not verified) :: Tue, 01/06/2009 - 2:29pm

@hector: "it's ludicrous to suggest that LaDainian Tomlinson was more valuable to the Chargers on Saturday night than Darren Sproles"

No one's suggesting this, and as noted above DYAR doesn't measure this. My point was "average, not outstanding." What DYAR does tell us is that, despite Sproles' amazing flash and game-winning TD, he didn't perform overall much better than the average NFL back in the situations in which he got the ball.

Barnwell's point here was spot-on; if the Chargers had lost the game, everyone and their brother would be piling on Sproles' fumble at the goal line as the critical play.

by hector :: Tue, 01/06/2009 - 5:53pm

hector: it's ludicrous to suggest that LaDainian Tomlinson was more valuable to the Chargers on Saturday night than Darren Sproles

jon: No one's suggesting this

The second table in the Quick Reads article reads "Five most valuable running backs." So, in my best John Madden, I'll say that "I always thought when you see a list of the five most valuable running backs, what you're really looking at is the five backs who give you the most value." Tell me where the disconnect is here.

by bengt (not verified) :: Wed, 01/07/2009 - 12:34pm

The disconnect is that the headline fits in the regular season, but not in the playoffs.

When you have five or more RBs at 50-100 DYAR, they are clearly the "Five most valuable running backs", irrespective of the way in which they gained their yards. When you have one RB at exactly replacement value (= 0 DYAR) after 20 carries and one RB who gains 1 DYAR with his only carry, they might both make it into the top five list in the postseason. But obviously the former is the more valuable player, because he has provided 20 carries of replacement value, which is preferable over one carry of replacement value plus one yard.

DYARs around 0 do not tell you anything about value just by themselves.

by Eddo :: Tue, 01/06/2009 - 1:30pm

Eh, I think Sproles's value is overrated by less analytical people. His 300+ total yards is inflated by return yardage, which DYAR doesn't measure. Return yardage is also easier to accumulate; every kickoff return is a guaranteed 20 yards, basically, and with the Colts pinned deep so often, Hunter Smith was more focused on getting off long punts, which are easier to return. Also remember, the Colts had very poor special teams coverage, so getting a lot of return yards is not as impressive.

by Jon Coit (not verified) :: Tue, 01/06/2009 - 12:37pm

There's also an error in the article, which references Sproles' "series of third-down conversions." He converted only one third down in the game.

by Love is like a bottle of gin (not verified) :: Wed, 01/07/2009 - 2:14pm

All together now, the numbers rate units NOT individuals. So Sproles in the SD offense, running the plays it ran, against IND D alignments was a -2. It is so contextualized it says very very little about Sproles himself.

by beargoggles :: Thu, 01/08/2009 - 12:11am

No comment on your comment except:
Great Handle!