Writers of Pro Football Prospectus 2008

11 Oct 2012

FO Mailbag: DVOA Explanation Triple-Header

Darin: Going into Week 5, the Redskins were ranked 10th (if I recall correctly) in third-down DVOA, despite being No. 32 by the league in "third-down efficiency." I chalked that up to too many third-and-longs, some of which were brought about by penalties. But penalties weren't a problem in Week 5, the Redskins were ineffective on third down ... and now are ranked No. 7?

Can you explain why DVOA is rating the Redskins so highly on third down?

Aaron Schatz: OK, so, first things first: opponent adjustments. Washington's VOA on third/fourth down is 5.9% (15th), as opposed to their DVOA of 17.6% (7th).

You have to remember that our "third down" efficiency is actually third or fourth down, and the Redskins have converted three of three fourth downs, so that gives them a little boost as well.

Another issue is the lack of turnovers. DVOA is not just measuring third downs as an either/or choice -- we are giving more credit for big plays, and a bigger penalty for a turnover that keeps your team from punting or trying for a fourth-down conversion or field goal. The Redskins have only one turnover on third down -- which actually isn't a turnover, because they recovered the fumble, but it's penalized by DVOA nonetheless. The average team has three turnovers on third down, and many of those are interceptions which are penalized a lot more than Robert Griffin's Week 4 fumble.

The flip of that, the Redskins also get a bonus for a couple of very nice big plays on third down, including the 77-yard touchdown to Santana Moss (the main reason their DVOA on third downs went up after Week 5). And even some of their "failures" get positive value over average because the situations were particularly difficult, such as a 14-yard gain on third-and-19 in Week 3.

I hope that helps explain.

Glenn Ferguson: I'm confused about the relative rankings of San Diego and Baltimore. If they have equal Pass Defense DVOA and SD has a better Run Defense DVOA, how does Baltimore have a much better Overall DVOA?

Aaron Schatz: This one's very simple, it has to do with play split. The Ravens have faced 192 passes, 164 runs. The Chargers have faced 215 passes, only 95 runs. So while San Diego has the better rating against the run, they've faced many more passes (where both teams are worse) and thus are worse overall.

Fred Tobin: I was watching the Giants' game last week and noticed at one point the following situation.

It was second-and-6, The Giants ran the ball, and Bradshaw appeared to pick up 2, maybe 2.5 yards in my mind and on the screen. However, the next play was listed as a third-and-3.

Even though it appeared to be closer to 3.5 yards and could have easily been listed as third-and-4, Bradshaw received credit for DVOA success on that play, as it got half the required yards to gain the first down on second down.

I was just wondering if you've looked into this in the past? The way the officials spot the ball and the stats pages list it, there are regularly yards to gain figures that may be a yard or two off.

Aaron Schatz: The answer is that "success points" in the DVOA system aren't actually binary. It's a curve. So except for when it comes to converting on third down, it isn't like there's a sudden jump in success that comes with one additional yard.

On the play in question, for example, Bradshaw gets 0.6 success points for the three-yard run on second-and-6. (The baseline for one full success point is 60% of needed yards, not 50%).

However, at a certain point we can't start watching the tape ourselves and trying to assign half-yards and quarter-yards. It becomes a time-sink to the point of ridiculousness.

Posted by: Aaron Schatz on 11 Oct 2012

11 comments, Last at 17 Jan 2013, 9:14am by http://www.curevaginaldryness.co.uk

Comments

1
by Dima (not verified) :: Thu, 10/11/2012 - 5:21pm

So the blurns are loaded, the count is three blurns and two anti-blurns, and the in field blurns rule is in effect. I think I'm starting to get the hang of this...

4
by dbostedo :: Thu, 10/11/2012 - 6:52pm

Except for the word "blern", that was complete gibberish.

5
by Malene, copenhagen (not verified) :: Thu, 10/11/2012 - 7:18pm

it's a futurama reference, dude.

6
by Malene, copenhagen (not verified) :: Thu, 10/11/2012 - 7:20pm

and so was yours, of course. I'm an idiot.

7
by dbostedo :: Fri, 10/12/2012 - 12:24pm

Maybe you're just in love...

Worm King: 'He's bluffing. No creature would willingly make an idiot out of itself.'
Fry: 'Obviously you've never been in love.'

8
by jebmak :: Fri, 10/12/2012 - 2:09pm

LOL

Also, goooooooooo Simpson's Paradox!

2
by Danish Denver-Fan :: Thu, 10/11/2012 - 5:25pm

The "a yard or two off" thing isn't much of a problem. The A in DVOA stands for average. So all players (and teams and units) are sompared too other players. Presumably Bradshaws bump is, on average, the same as his peers'.

3
by Brian Anderson (not verified) :: Thu, 10/11/2012 - 5:48pm

The second question is an example of Simpson's paradox:

http://en.wikipedia.org/wiki/Simpson's_paradox

9
by rmgbd.org (not verified) :: Sat, 10/13/2012 - 11:14am

Woω, wonderful blοg fоrmat!

How long hаvе you been гunnіng a blog foг?
you maԁe runnіng a blog lοоκ easy.
Thе entirе look of youг ωeb site is magnіficent, aѕ well aѕ the content materіal!

10
by Julia (not verified) :: Sun, 10/14/2012 - 7:58am

Hаνe you ever thοught about ωriting an еbook or
guest аuthoring on otheг webѕitеs?
I haѵe а blog bаsed upon οn the same subϳects you disсuss and wоuld reallу like tο have you shaгe ѕome
stories/information. I know my гeadеrѕ wοuld vаlue your wοrk.
If you are even remotely іntereѕteԁ, feel frее to ѕend me an emаil.

11
by http://www.curevaginaldryness.co.uk (not verified) :: Thu, 01/17/2013 - 9:14am

Hello there, ϳust beсame awaгe of уour blog
thrοugh Googlе, and found that іt iѕ гeаlly informаtіve.

I'm gonna watch out for brussels. I'll appreciate if you сontinuе this іn future.
Lоtѕ of peοple wіll be benefited from youг ωгiting.
Cheers!