My personal case against DVOA (A hypothetical situation with made up numbers driven to an extreme)

[shaka]

I was lurking on that Seahawks abomination of a message board when someone had said something along the lines of, "well you Panther fans just don't like DVOA because it's complicated.  Just read about it, it's not that bad."  So I did.  

 

DVOA - Defense-adjusted Value Over Average.  Part of the explanation from the DVOA site:

"Going back to our example of the three-yard rush, if Player A gains three yards under a set of circumstances in which the average NFL running back gains only one yard, then Player A has a certain amount of value above others at his position. Likewise, if Player B gains three yards on a play on which, under similar circumstances, an average NFL back gains four yards, that Player B has negative value relative to others at his position. Once we make all our adjustments, we can evaluate the difference between this player’s rate of success and the expected success rate of an average running back in the same situation (or between the opposing defense and the average defense in the same situation, etc.). Add up every play by a certain team or player, divide by the total of the various baselines* for success in all those situations, and you get VOA, or Value Over Average."
 

There, I bolded the important part for you.  Now let's take a hypothetical look at why anything Value Over Average should be taken with a grain of salt.  Warning: numbers, math and a bunch of hypothetical bullshit starts now.  Yes, there's a ton of cherry picking and a lot of oversimplification.  I don't care about you trying to punch holes in my argument, this is my take on DVOA and why it's not worth my time.  I will not respond.  I give zero fugs.  There are zero fugs given.  

There were 6845 third downs in the NFL during the regular season this year.  Of those 2667 were converted.  For the sake of easier math, let's round those up.  

7000 3rd down plays.  

3000 conversions.  

 

This is where I start making things up to illustrate an extreme.  Let's break it down further, assuming a uniform distribution on third down distances and only looking at the converted plays.  

1000 Third and short (less than 3)

1000 Third and medium (4-7)

1000 Third and long (8+)

 

Now lets break down third and short into the following:

500 Third and one

500 Third and two

 

And now third and one plays by conversion play.  

50 Penalty

300 Run

150 Pass

 

Now let's look at third and one routes ran

Go/Seam - 30

Curl - 60

Out - 60

 

Further, lets look at just the seam routes

10 went for touchdowns, 20 did not

 

Now let's look at the 20 that didn't go for touchdowns because arguably plays that go for touchdowns should be worth more.  (15 numbers generated randomly between 7 and 40, 5 numbers generated randomly between 30 and 75)

9,12,13,13,15,19,23,23,25,26,29,34,36,37,39,42,55,61,65,72

Average of:  32.4 yards 

 

Now with this knowledge, let me present a scenario.  It's the first quarter.  The Panthers win the toss and elect to defer.  The Saints have a 9 play 75 yard 7 minute drive before the defense forces a turnover at the Carolina 5 yard line.  The Panthers ran the ball twice, and are in a third and one on our own 14 yard line.  Greg Olsen is bracketed on the play, runs a seam route and makes a one handed catch for 9 yards.  Let's look at the VOA:

9 yards compared to the 32.4 yard average on third and one seam routes.  Look at this piece of poo Greg Olsen, he isn't even remotely close to average.    

But wait, now throw in the fact that the Saints are a historically bad defense.  The Saints on average give up 11 yards on third down because they're assholes.  Jesus Christ Greg, you can't even get 11 yards on third down.  Relative to other third downs against the Saints, that's a below average play.  

Facts:  9 yard catch on third and one.  

Positives:  converted on third down

Negatives:  Well below average in similar circumstances.  Below average against the defense

 

Obviously the real DVOA is a lot more refined.  But you can begin to see the cracks in the metric.  I don't give a flying poo about how we performed relative to the NFL average when the NFL average still contains the Browns.  I feel like we're still missing something...

 

Oh that's right, the actual football game.  First of all, one handed catch while being doubled.  That's a hell of a catch and is in no way, shape, or form average.  Secondly, the value of picking up the first down in this situation is gigantic.  I don't care if this possession results in points.  Give me a first down or two, let my defense catch their breath and try to get the Saints offense out of rhythm.  You know moleface and vicodin boy are kicking themselves for not getting points, and want that ball back bad.  The best defense is keeping their offense off the field.  The Saints historically bad defense works against us in this situation, but where's the weight or emphasis on the third down conversion that keeps the #1 passing attack on the sideline.  It's an interesting take on traditional stats that is decent but just feels..incomplete.  

 

TLDR:  DVOA is a good metric, and is better than your traditional stats.  However, there's a lot more to football than performing against the average.  Where's the metric that accounts for going back to a wide receiver on a crucial third down when two plays earlier he dropped a long touchdown?  Or the metric showing the impact of a physical running game on pass defenses over time?  What about pass velocity by route and the correlation to drops?  And penalties committed but not called by refs?  Successful play types by defense called?  How much of a sample size will you really get from third and short go routes against double a gap blitzes?  It sure as hell won't be credible.  

 

The game of football is complex, and there will never be a metric to account for every little thing that happens during a play.  

 

Summary of the TLDR:  I used to ignore DVOA.  I still ignore DVOA.  fug the Seahawks.  

[PandaPancake]

I'll read it fully later because you put in such work and I like to fact check. But I wanted to remind you of this: "The difference between an engineer and a engineering student? The student tells you the answer and the engineer asks, 'What do you want the answer to be?"

[stankowalski]

 

[FootballMaestro]

I don't have a problem with DVOA and other analytics. 

It's just one tool. Nothing else. That's it.

Problem with many Seattle fans: They only know analytics (talking about it so much), cause many of them are probably too lazy, and/or dumb to watch film and other players, teams.

 

[ecu88]

1 hour ago, PandaPancake said:

I'll read it fully later because you put in such work and I like to fact check. But I wanted to remind you of this: "The difference between an engineer and a engineering student? The student tells you the answer and the engineer asks, 'What do you want the answer to be?"

Same with economists:

An accountant, a mathematician, and economist interview for the same job. The interviewer asked the accountant what does 1+1 equal? The accountant answered within the percentages, the answer is 2.  Next was the mathematician interviewed, asked the same question he responded. Through validated theory 1+1 equals 2. Finally the economist was interviewed.  The interviewer asked the same question to the economist as he did for the others. The economist stood up closed the door behind, locked the door and pulled down the blinds then said, what do you want it to be.

[PandaPancake]

Woah, i think we Pi'd together

[Cary Kollins]

yep, made up stats gonna made up stat

 

Give me the 15-1 team with the MVP and 10 pro bowlers.

[Quinning]

1 hour ago, shaka said:

Summary of the TLDR:  I used to ignore DVOA.  I still ignore DVOA.  fug the Seahawks.  

... and if you look at the top 20 DVOA's of all time.... 4 in the last 30 years won the Super Bowl.  ('91 Redskins, '85 Bears, '96 Packers, '13 Seahawks)

They've already crowned themselves...   Their fans reek of arrogance.

End their season on Sunday. 

37 years in Georgia... the last 4 in Seattle.  I didn't become a 2012th man... sorry, not interested.

Beat them and we all get another year of not having to listen to them.

Cheers.

 

 

[frash.exe]

If the Panthers were better at DVOA, you wouldn't hear Seahawks fans talking about this. It's not DVOA they value, it's their team and they want to find any reason they can hang their hat on to reaffirm what they think.

on Sunday, it's not going to matter. Either the Panthers are going to win or the Seahawks are. DVOA does not determine the outcome of football games in the future. 

[t96]

2 hours ago, Quinning said:

... and if you look at the top 20 DVOA's of all time.... 4 in the last 30 years won the Super Bowl.  ('91 Redskins, '85 Bears, '96 Packers, '13 Seahawks)

They've already crowned themselves...   Their fans reek of arrogance.

End their season on Sunday. 

37 years in Georgia... the last 4 in Seattle.  I didn't become a 2012th man... sorry, not interested.

Beat them and we all get another year of not having to listen to them.

Cheers.

 

 

Are you a Falcons fan in Seattle who hates the Seaturkeys more than the Panthers and is actually actively cheering for the Panthers and against the Seaturkeys this week?

[KB_fan]

15 minutes ago, frash.exe said:

If the Panthers were better at DVOA, you wouldn't hear Seahawks fans talking about this. It's not DVOA they value, it's their team and they want to find any reason they can hang their hat on to reaffirm what they think.

on Sunday, it's not going to matter. Either the Panthers are going to win or the Seahawks are. DVOA does not determine the outcome of football games in the future. 

Agree with this.  I try to avoid using stats as an attempt to predict the game.  I find them much more interesting in terms of description and analysis looking back - (what helped shape the outcome of a game?) - or looking over time to see how a team is improving or declining over a span of time.  Yes, rankings are interesting, but not the be-all end-all, as we all know, for instance re: the QB rankings that focus solely on completion %.  That metric alone could not encompass all that Cam was bringing to the game and leading us to wins.

DVOA is somewhat interesting, and I've enjoyed watching the Panthers' DVOA rankings climb throughout the season, but trying to turn it into a prediction tool is something I find silly.  

Teams are made up of individuals and there is also team chemistry.  Stats & metrics like DVOA predicted that Panthers should only have gone 12-4 this season.  What's the extra difference that helped us "overachieve" and win 3 more games than predicted?  (By the way, the highest "overachievement" in terms of actual wins vs expected wins of any team in the league this year.  See here:  http://www.businessinsider.com/carolina-panthers-have-been-lucky-2015-12)

All the stats predicted we would beat Atlanta in the 2nd game, but we lost.  A dropped pass here or there.  Cam and the team looking a bit flat post-Christmas...,  Stats don't account for that.  This game is played by people, not computers, so statistical analysis can only go so far.

[Quinning]

4 minutes ago, thomas96 said:

Are you a Falcons fan in Seattle who hates the Seaturkeys more than the Panthers and is actually actively cheering for the Panthers and against the Seaturkeys this week?

That'd be correct...  but TBH there isn't really any hate here towards the Panthers.

The Saints are MUCH more hated than Carolina ever has been...  Seattle on the other hand... they top the list.  Their arrogant fans are to blame.  I hear it on a daily basis.  The dumpster fire that was their team for the first half of the season had their fans in hiding... the 2nd half of the season brought them out for all the holes they were hiding in... and the win over Arizona in week 17 scribed their name on the SB 50 trophy.

That's what their fans say anyway...