Writers of Pro Football Prospectus 2008

16 Dec 2007

ESPN: Vikings' Run Dominance Might Not Be Enough

This week's ESPN MNF article discusses the myth that run-oriented teams succeed in the postseason, particularly where it concerns this year's Minnesota Vikings -- one of the most unbalanced teams in NFL history, with offense and defense far superior on running plays as opposed to passing plays.

Posted by: Aaron Schatz on 16 Dec 2007

28 comments, Last at 19 Dec 2007, 5:24pm by BDC

Comments

1
by jebmak (not verified) :: Sun, 12/16/2007 - 2:16am

Aaron Schatz is president of Football Outsiders Inc.

So you guys incorporated eh?

Yep, a whole football article, and that is what I get out of it.

2
by QB (not verified) :: Sun, 12/16/2007 - 2:19am

I could have sworn we saw a study last year showing that although the passing game is generally what decides games in the regular season, postseason success has correlated most strongly with running/defending the run.

Anybody else remember?

3
by Temo (not verified) :: Sun, 12/16/2007 - 2:45am

2. That might be the case, I don't know... but in any case that wasn't exactly Aaron's point here. He was looking into the performance of teams that are both dominant in the ground and bad in the passing game. You're talking about a related yet different idea.

4
by Bobman (not verified) :: Sun, 12/16/2007 - 2:54am

#2 I don't know, that sounds an awful lot like standard conventional wisdom and/or a cliche (especially with outdoor games in cold climates emphasizing the run game)... and FO is just the place to turn it upside down. (Especially with more domes and southern teams now compared to 1970 for example).

I don't remember that article, but the turnaround of the Colts D in the playoffs might be a good place to look. Clearly they won, largely because of their stiff D, and that was largely a surprising run-stopping D. So if that article popped up last year, that's a logical place to look.

But I'm thinking no.

5
by Yaguar (not verified) :: Sun, 12/16/2007 - 3:08am

I think Bill might have done a study showing that some things seem to account for greater success in the playoffs than in the regular season, and vice versa. But I don't remember it all that well, and I think it's difficult to do a good study of that.

6
by Eddo (not verified) :: Sun, 12/16/2007 - 3:20am

2,4,5:
Bill Barnwell did write up a study last year, right before the playoffs, called "Why Doesn't Bill Polian's S--t Work in the Playoffs?" (it also appeared in a form in PFP '07). Basically, it did show that a teams defensive rushing DVOA correlates best with postseason success*.
(* - Although this is possibly my favorite article to appear on FO (it provoked a mini study of my own), Bill's definition of playoff success was kind of arbitrary. I've linked the article if anyone is more interested.)

7
by BDC (not verified) :: Sun, 12/16/2007 - 4:06am

"Conventional wisdom says that to win in the playoffs, you have to be able to run the ball and stop the run."

And conventional wisdom would be correct. Lets look at the past few SB winners, shall we?

INDY: 6th in rushing offense, 31st in defense. But we all know they got much better in the playoffs on D, particularly, rushing D. Their opponent, Chicago was 9th and 5th.

PITT: 10th in rushing, 1st in stopping the run. Their opponent, SEA, was 3rd and 3rd.

NE: 4th in rushing, 6th in stopping the run. Their opponent, PHI, was 6th and 15th.

NE: 24th in rushing, 9th in stopping it. Their opponent, CAR, was 20th and 13th.

TB: 27th in rushing, 7th in stopping the run. Their opponent, OAK, was 4th and 4th.

NE: 18th and 26th, widely regarded as one of the weakest SB winners. Their opponent, STL, was 1 and 7.

BAL: 7 and 1, their opponent, NYG, 24 and 4.

STL: 3 and 6, their opponent TENN 4 and 4.

DEN: 1 and 9, their opponent ATL 8 and 2.

As you can see, almost all of these teams were strong in either running, or stopping the run, and most of them were strong in both.

What does this mean as far as MIN goes? Well unfortunately for MIN, "Conventional wisdom says that to win in the playoffs, you have to be able to run the ball and stop the run." It does NOT say you have to be strong in those areas at the *expense* of sucking in the passing game. Most of the above teams were good at passing, good at stopping the pass, or both. That does not bode well for MIN, as they aren't particularly good at either.

8
by Dan (not verified) :: Sun, 12/16/2007 - 4:39am

Hey its said the you actually make money perhaps doing this.
In medicine we remember numbers. Like 28 or 29 for example. Sad. I hope your precious website isn't all you have.

9
by Anthony (not verified) :: Sun, 12/16/2007 - 6:38am

This is an interesting idea, but most (though not all) of those teams were bad because they were bad in the running game and the passing game - just a lot worse in the passing game.

Since the Vikings ranks in the 4 categories explored (run offense, run defense, pass defense and pass offense) are 1, 2, 15 and 20 they have a "total score" of 38. We reach this by adding the scores for all 4 categories, so this “total score” represents the sum of a team’s ranks in each aspect of play from scrimmage.

The "total scores" of the other teams listed in this article range from 54 to 90. The lower a team’s “total score,” the better – it means that team ranked closer to number 1 in more categories. [It may be no coincidence that the team with a total of 54, the 1984 L.A. Rams, made the playoffs with a record of 10-6.]

As you can see, the Vikings (38) – taken as a whole – rank significantly better than any of these other teams (54-90).

This evidence allows me to put forward a hypothesis (though I certainly would not claim to have proven it):

Most of the teams examined in this article were extremely bad in the passing game, which is what allowed them to do so much better in the running game than the passing game.

A team that is horrible passing has a relatively easy time being much better running than passing.

It works vice versa as well: a team that is horrible running has an easier time being much better passing than running – but it is still not likely to be a good team.

Thus, this article does not show that a team that is good in the passing game but bad in the running game is in better shape for the postseason than a team that is good in the running game but bad in the passing game (which may be true, but this article does not prove that).

This article shows, if anything, that a team must be good in both the running game and the passing game in order to succeed in the postseason. The statistics put forward do not justify any other conclusion.

This is still a very interesting article - I'm not trying to tear the author down at all. The thinking is very original and it raises good points that should be investigated further.

This article also shows the importance of run-pass balance: to win, it helps to be good in all phases of the game.

And, for the record, I do agree with the final verdict – I predict the Vikings will lose at Seattle in the wildcard round.

+++++++++++

“Total Scores”:

2006 Atlanta: 68
1984 L.A. Rams: 54
1999 Cincinnati: 66
2000 Cincinnati: 65
1999 San Francisco: 65
1998 Pittsburgh: 63
1983 Pittsburgh: 64
2007 Baltimore: 81
1975 Cleveland: 59
2005 San Francisco: 90
1985 Indy: 58

10
by Anthony (not verified) :: Sun, 12/16/2007 - 6:42am

Also, just wanted to second BDC - just being good at the run game isn't enough. A team has to be good at passing as well to have the best shot at advancing in the playoffs.

11
by ammek (not verified) :: Sun, 12/16/2007 - 7:01am

What about the champions that were generally considered to be run-heavy? I'm thinking: 72-3 Dolphins, 74-5 Steelers, 82 Redskins, 92-5 Cowboys and 97-8 Broncos. Could you give us their rankings? I'm guessing they were league-leaders on the ground, with a decent passing game to surprise defenses when they needed - kind of the opposite of the 2007 Pats.

12
by Tom (not verified) :: Sun, 12/16/2007 - 7:31am

Re 9:

Luckily we have a stat which combines both running and passing while remove luck. This happens to be DVOA, in which the Viking rank 10, which is pretty good.

13
by MC2 (not verified) :: Sun, 12/16/2007 - 10:01am

I've always been suspicious of these "magic formula" theories about postseason success, e.g. "good pitching beats good hitting", "defense wins championships", etc. At the risk of stating the obvious, it seems that most championship teams are the teams that are the most balanced, with the fewest flaws, rather than those who do just one or two things extremely well. Of course, there are always going to be exceptions, and whenever they occur, it generally leads to the talking heads proclaiming that Team X has just laid forth a "new blueprint" for postseason success. Funny how those blueprints rarely seem to work for more than a year or two, if that.

14
by jimm (not verified) :: Sun, 12/16/2007 - 11:02am

The article simply supports what is already known - passing and stopping passing efficiently is the key to winning. However, with respect to the Minnesota Vikings the question is are they really just average with respect to passing. Over the whole season they are but in their last 6 games they've dominated the passing game in 5.

Pass yds per play differential Minn stat first opp avg second.

SD 6.5-4.3 = +2.2
GB 5.6-7.7 = -2.1
Oak 9.3-7.0 = +2.3
NYG 7.8-4.5 = +3.3
Det 8.4-5.9 = +2.5
SF 6.5-4.3 = +2.2

15
by fiddycentbeer (not verified) :: Sun, 12/16/2007 - 12:55pm

This article debunks a theory which doesn't need debunking. With the advent of the Manning Rules (or Manning Emphasis if you prefer), that hoary paradigm, run to set up pass, has been inverted. Everyone knows that.

Ground power is a fun index, but not an especially useful one. First, it elevates the utterly helpless, like those ATL teams. Second, as any student of situational football knows, YPC is not a useful stat. Subbing in a RB carries differential for YPC, while retaining YPA, may yield something more useful. I dunno, but here's one e.g.

The 2004 Pittsburgh Steelers went 15-1, classically using a balanced attack to get a lead H1 and then grinding time thru H2. That team finished #1 in RA and, iirc, #1 (as in fewest RB carries against). However, they were 19th and 5th as to YPC. Since they were 3rd in YPA, each way, those 2004 Pittsburgh Panzers show a negative GP, as constituted.

That does not describe the facts. I appreciate that this is simply one example, and that those Steelers were outlying wrt RA. Still, the article's fatal flaw remains: YPA means bupkus.

16
by fiddycentbeer (not verified) :: Sun, 12/16/2007 - 12:57pm

Oops.

Sorry. Last line should read: YPC means...

FWIW, i'm big on YPA.

17
by mmm... sacrilicious (not verified) :: Sun, 12/16/2007 - 7:31pm

There is substantial confusion here between correlation and causation.

When your team is losing, you pass more. When you pass more, the defense expects pass more. When the defense expects pass more, rushing YPA go up.

Conversely, when your team is winning, you run more, the defense expects run more, and rush YPA go down.

I would be willing to bet that this phenomenon explains a lot of this "paradox".

18
by Andrew (not verified) :: Mon, 12/17/2007 - 12:04pm

Any layman can make some simple correlations too using the database on profootballreference.com. In the 16 game season and modern passing rules since 1978 (and excluding the strike seasons of 1982 and 1987), here's some r^2 values comparing various conventional metrics to pythagorean wins.

points for: 0.6243
points against: -0.5619

yd/att: 0.3111
yd/rush: 0.0438

total TD: 0.4659
offensive INT: -0.1654

rush yds: 0.1912
pass yds: 0.0714
total yds: 0.2967

completion %: 0.1623
pass/run ratio: -0.1181

pass yards against: -0.0043
rush yds against: -0.2880

yd/att against: -0.2183
yd/rush against: -0.0584

I don't think this is telling anyone anything they don't know except perhaps that passing yards correlate out as less important that yards per attempt, while yards per rush (and against) is far less important than rushing yards and rushing yards against.

Winning teams have efficient passing offenses and defenses and overwhelming (but not necessarily efficient) rushing offense and defense.

DVOA (and watching games) tells us that the overall team efficiency in getting leads is what causes the overwhelming rushing game, because having a lead allows the offense to drain the clock with the run and forces the opponent into passing to catch up, where interceptions are much more likely than turnovers.

The important thing appears to be playing to whatever is your strength that can give you a lead, and then draining the clock through the run to secure victory.

For Minnesota from the games I've watched, their strengths are not just rushing and stopping the run, but also a very opportunistic defense that puts a lot of points on the board through interceptions and interception return TD's, and a developing efficient passing game that supplements their main strength in running. If anyone bothered to check the stats, they'd see we have not seen a defense score this many points in a long time (the 99 Chiefs and Rams both had 11 def/ST TD's in a full season). They have 9 defensive/special teams touchdowns in 13 games!

Since this is a team that was a blocked field goal returned for a TD away from taking Dallas to the wall, I think they can hang with anyone in the NFC. They've certainly shown they can get on a roll and win some games by really big margins through their type of play. Plus Coach Childress has plenty of playoff experience with the Eagles. If Dallas or someone else dismisses them lightly pre-game as Dallas did this week with the Eagles - watch out!

19
by Brian (not verified) :: Mon, 12/17/2007 - 1:00pm

17 and 18 hit the nails on the head here.

There is a difference between offensive imbalance and defensive imbalance.

If an offense is imbalanced it's not as harmful because the offense can choose to emphasize their strength.

An imbalance on defense is probably far more fatal because the opponent can choose to attack the weakness.

20
by BD (not verified) :: Mon, 12/17/2007 - 1:01pm

Re: 12

If a measurement (DOVA) is made up of 2 constructs (rushing and passing), and these two constructs are significantly different for an individual (i.e. rushing much greater than passing) then usually we say in the measurement world that the overall measurement (overall DVOA) is less meaningful.

So, for example, if we are testing the IQ of a child who has brilliant verbal skills but weak visual skills, that child's overall IQ (which would be average) doesn't really tell me anything about that child.

In this case, for Minnesota on defense, their rushing DVOA is -30.0, while passing is 10.2. In other words, they can stop the run but not the pass. Their overall DVOA is -4.8, which does not seem to adequately describe their defense (except to say that perhaps stopping the pass is more important).

I'm not sure at what point we would say that two DVOA measures are significantly different, but I believe that the 40-point difference in Minnesota's scores definitely qualify. In this case, overall defensive DVOA would not be a good measure of how successful their defense is.

I hope that made sense without being too technical.

21
by Devin McCullen (not verified) :: Mon, 12/17/2007 - 1:09pm

I'd be interested to see the results of the reverse teams, the "air power" index. Other than last year's Colts, I doubt they'd be any more successful than the ground power teams, but it would be nice to know.

22
by Methodologically Flawed (not verified) :: Mon, 12/17/2007 - 3:47pm

Umm, this article seems only to show that teams that are severely unbalanced (insofar as they are really good at some things but really bad at others) don't do well in the postseason.

To that, I would politely reply, no $hit.

The appropriate counterpoint here would be to compare the results of strong "run power" teams with strong "air power" teams, i.e. teams that really suck at defending the rush, or running the ball. Instead, I see rhetorical references to the efficiency of passing and stopping the pass.

Also, to point out that "run power" teams play better in the regular season than the postseason is also a "no $hit" point. This point is in fact true for ALL playoff teams as a whole, since their competition is better.

I would hazard a guess that teams with a tilt towards "run power" outperform, relatively speaking, teams with a tilt towards "air power" in the wildcard, 1st round and conference championship rounds of the playoffs, for the same reasons that passing statistics tend to drop off during more inclement weather.

23
by Scott C. (not verified) :: Mon, 12/17/2007 - 6:18pm

#17. You postulate that teams that get ahead and run more end up with lower yards per rush attempt because the defense is expecting it. I don't think thats generally true.

For some teams, or in blowouts, perhaps. But I'm fairly sure that over the average of all games where a team has more than 20 rushing attempts and leads at halftime, the yards per carry INCREASE each quarter. Look at Ladanian Tomlinson's splits by quarter -- Defenses get tired and most of the long runs occur in the second half.

Clearly, not all teams can "wear down" opposing defenses with the run, but its a well known phenomenon you conveniently neglect.

24
by Joe T. (not verified) :: Mon, 12/17/2007 - 6:24pm

Not sold on the "ground power" stat. In fact, I hasten to call it a stat since its actually a difference between rankings, which are not really hard "stats" anyway.

Maybe there are some fair assumptions and theories in the article but calculating "ground power" doesn't really give us the tools to test the hypothesis.

IMO in seems like most defenses are configured to defend the pass, hence the proliferation of the Cover & Tampa 2 throughout the league. I suspect that we get the assumption that the ground game wins in the playoffs because teams that torch lesser defenses through the air in the regular season are of course going to struggle against stronger defenses in the playoffs and be forced to resort to the run for yardage and to open up their passing attacks.

25
by Joe T. (not verified) :: Mon, 12/17/2007 - 6:37pm

#24 by hasten i meant hesitate - not enough caffeine today

26
by gmc (not verified) :: Tue, 12/18/2007 - 9:24pm

re: #7

So... Good teams tend to play in and win Super Bowls? Who knew?

No, I'm not knocking the point, just trying to get a laugh. Thing is, unless you do what Aaron did, and compare ground dominance to air dominance, you'll effectively capture success through the air while crediting the effetive ground game. St. Louis in 1999 and 2001 wasn't really a -rushing- team were they.

Aaron's formula has a problem too, in that teams which run the ball often tend to be less effective in doing so the more they do it. Exhibit A: The Eagles amazing DVOA stats, which are far more related to the fact that they only run one running play a quarter than the Philadelphia offensive line.

27
by cjfarls (not verified) :: Wed, 12/19/2007 - 4:54pm

I tend to agree... I found the analysis of this article kind of lacking, as the core stat used (ground power) really doesn't tell you anything. Comparing "ground power" to "airpower" or something of that nature might have been interesting, but simply saying "teams that are bad at passing & pass defense are unlikely to do well in the playoffs" leaves me saying "well, duh!". Does that mean teams that are bad at running and run D are GOOD in the playoffs... my guess is no as well.

As many folks have said, in the playoffs, any weakness in a team is likely to be exploited. I think that is true for pass D, run D, special teams, air attack, run attack, etc. You really need to be at least average everywhere, or you'll get killed.

The Vikes this year have been amazing running, because really no one has been able to shut down the ground game, even when selling out completely against the run. But even given that, their pass Def has been average and the pass Off maybe improving. I don't trust their pass Offense to bail 'em out against a good run D, but thats why no one considers them one of the "elite" teams.

Would anyone be surprised if they went one and out? No... which makes this article kinda dumb for predicting such.

28
by BDC (not verified) :: Wed, 12/19/2007 - 5:24pm

26: I don't know, I seem to remember some guy named Faulk being pretty good :)

All joking aside though, yes, I know it seems obvious that good all-around teams win SBs, but hey, I'm not the one that wrote the article questioning that!

My point, which I tried to make clear, isn't that running the ball or being able to stop the run is MORE important then the pass, but that it is still important. If your team is weak defensively against one of those two, the other team can exploit your weakness, provided that they themselves are strong in both; if they are weak in that category you are weak in, then they can't.

We all know how important the pass is in the current NFL; it doesn't really need to be mentioned. The run on the other hand is sometimes forgotten. That is the only reason for mentioning it. Not because it is more important then the pass (quite the opposite), but simply because it does still have a role. I tend to think it more like, to be successful in post season, a prerequisite is to be good at the pass, but that being good at the run can provide that extra something to win the SB.