Sunday, August 10, 2014

Email Bag: Running to Win or Running after Win?

I love advanced metrics.  I love statistics.  I love thinking about new ways to strike down old myths.

I love all of these things, so what you are about to read may be problematic.

In baseball, if I read something from Fangraphs, I generally am not fact-checking.  I take them at their word because they immerse themselves in baseball all summer long.  I don't.  I immerse myself in baseball for 3 hours at a time, but certainly not for weeks or months at a time like those dudes.

But, football?  That's my jam.

So, I want to read your theories and your premises, but they have to pass my tests of observation and study.  I am watching football nearly 12 months a year and almost every day inside those 12 months in some form or fashion.

I want to hear your statistical study, but I am also attending a coaching clinic, talking to scouts, and watching film on my own.  

And if it all fits, great.  But, if the stat guys are saying something that doesn't agree with the coaches and scouts, then someone needs to say something.

You see, statistics can tell us plenty, but it doesn't answer "why?" very often.  Why do we run?  Why do we pass?  Why do we play this guy in this situation and then this guy in this situation?  DVOA never answers that.  They simply measure results and compare numbers.

So, the numbers side of football and the tactical side of football live in two separate universes.

But, what if a guy like me likes to live in both?

Do I silently disagree when I read things that don't hold water with the other side?  This must be why the stat guys and the scouts were always mad at each other in Moneyball.  The stat guys have it all figured out and think that observation only gets in the way sometimes.  

Honestly, when it comes to my football advanced metric thought leaders, Football Outsiders are my guys.  I buy their annual every year.  I love it.  I dive into it.  But, the cornerstone of everything they write and believe is based on the following essay which starts every annual:

You run when you win, not win when you run.
 The first article ever written for Football Outsiders was devoted to debunking the myth of "establishing the run." There is no correlation whatsoever between giving your running backs a lot of carries early in the game and winning the game. Just running the ball is not going to help a team score; it has to run successfully.
There are two reasons why nearly every beat writer and television analyst still repeats the tired old-school mantra that "establishing the run" is the secret to winning football games. The first problem is confusing cause and effect. There are exceptions, usually when the opponent is strong in every area except run defense, like last year's Pittsburgh Steelers. However, in general, winning teams have a lot of carries because their running backs are running out the clock at the end of wins, not because they are running wild early in games. 
The second problem is history. Most of the current crop of NFL analysts came of age or actually played the game during the 1970s. They believe that the run-heavy game of that decade is how football is meant to be, and today's pass-first game is an aberration. As we addressed in an essay in Pro Football Prospectus 2007 about the history of NFL stats, it was actually the game of the 1970s that was the aberration. The seventies were far more slanted towards the run than any era since the arrival of Paul Brown, Otto Graham, and the Cleveland Browns in 1946. Optimal strategies from 1974 are not optimal strategies for today's game. 
A sister statement to "you have to establish the run" is "team X is 5-1 when running back John Doe runs for at least 100 yards." Unless John Doe is ripping off six-yard gains Jamaal Charles-style, the team isn't winning because of his 100-yard games. He's putting up 100-yard games because his team is winning.

Every time I read that, I cringe.  Because they write it in such a manner that it basically relegates all old football people as being confused.  They are so cute and don't understand our method of arithmetic and therefore allow us to explain their game to them.  

What makes it worse is that, on the surface, it makes sense.  We all have memories of our favorite RB being given the ball in the 4th Quarter of a game and the defense can't get a stop.  Emmitt Smith was running the ball down Green Bay's throat in the 1995 NFC Championship Game, and they couldn't stop him.  He was running because the Cowboys were winning.  He was the closer, picking up the save by killing the remainder of the clock.

But, it doesn't even address all of the countless tactical/strategic reasons for running that play out Sunday after Sunday.  Then, countless Madden and Fantasy Football-playing fans read this and agree, partly because it matches their sensibilities for what works in both Madden and Fantasy Football.  It must be right.  Forgive me if I am assuming Head Coaches or Offensive Coorinators don't read Football Outsiders as much as fans, but I am taking that leap without doing the research.

That, of course, leads to people assuming that Football Outsiders have this thing nailed (after all, they tell us that it is the first story they have ever written).  That, in turn, brings it to readers quoting it as if it was sent down from Heaven above.  Gospel.  It must be.  Running is pointless and despite coaches believing in it for a myriad of reasons, we now have proof that it doesn't correlate to winning at all.

Now, to debunk their premise, I could offer you an extend film session on the actual effects of running the ball and physical football.  The coaches and scouts tell you repeatedly that their are two ways to defeat an opponent.  Speed and Power.  Speed is certainly more conducive to highlights and video games, but power still works and always will.  Chip Kelly loves speed and Jim Harbaugh loves power.  The NFL evolves around and around, but at some point you are talking about speed and power or some combination of the two.  The comedy of this, of course, is that Chip Kelly loves speed, but does so inside a running game that is often 60% run at Oregon.  Harbaugh loves to pass, too.  Balance is what we seek.

Quick thought: When I was recently at a coaching clinic, I asked one HS coach about this theory and where he sat on "establishing the run".  His entire reasoning is based on the following - do you think attacking a castle is easier when you surprise them or when they are waiting for you?  You want to pass when they are waiting on the run, not the pass.  It is just trying to enhance your chances of success.

But, really, I think anyone who has ever watched a game knows this. Where I agree with Football Outsiders is the claim that giving Emmitt the ball 20 times is the key to winning as some magical number or formula for victory.  We used to joke that we might as well just take a knee 20 times and presto!

Of course those stats are misleading, especially if most of the runs come after the lead is established.  But the effects of running the ball are almost countless:

* - Running the football forces the defensive front into physical exhaustion, Passing the football forces the offensive line into physical exhaustion.  Are you attacking or defending?  If you are pass protecting, you are in retreat against an attacker.  If you are run blocking, you fire forward at the snap and the defense is trying to defend their spot.  Exhaustion, of course, leads to injuries.  Think of all of the OL injuries.  I suspect most in your head are from pass protection.  DL injuries?  Most in my head are from run stopping, not pass rushing.

* - Running the football slows down pass rushers.  It also stops blitzing.  Catch the defense in a blitz with a nice draw?  That burns your hand and you will be less likely to blitz.  But, why not blitz when you know they are passing all of the time?  Pass rushers also have run responsibilities, but if you oblige them by not running, they become nothing but "rush men".

* - Running the ball keeps you ahead of the sticks.  Even an unsuccessful run play generally gets 2 yards.  Do you like 3rd and 10 or 3rd and 6?  Well, incompletions get nothing and interceptions are worse.  Runs have a lower ceiling, in general, but a higher floor.

* - Perhaps most importantly, Running the ball allows you to control the pace of the game.  If your defense has issues, you can shorten the game.  You can protect the defense by controlling the number of snaps they must stand up to a test by the opposing offense.  If every game averages 11 possessions to each side, would you enhance your situation by limiting each to 8 or 9?  On the other hand, what happens if your pass-happy 3 and outs put Denver or New Orleans back out there for 14?  If the best defense is a good offense, then much of that cliche has to be based on controlling the clock and time of possession.

* - Running the ball occupies the Linebackers and Safeties in their pre snap reading of keys.  This is the beauty of play action passing.  Balance leads to indecision and confusion for the defense.  Playing "Madden-style" (80-90% passing) leads to no confusion whatsoever.  So, I am saying it - running the ball makes passing the ball so much easier.  

All plays are organically connected in football.  Tendencies are the basis for game plans.  Players are calculating what is about to happen in ever huddle.  They rely on what you just did to plan for what you are about to do.  All of this is connected.  It isn't about getting a "magical 20th carry".  It is about confusion and misdirection to get a defense out of position for that next run or pass that might win the game.

And back to Football Outsiders, it isn't even about attacking a weak run defense.  It is about making a defense defend everything.  About making indecision their biggest enemy.  That's all.  

I love those boys, over there.  But, I think their simplicity in this matter is too big of a departure from the logic coaches and scouts use in real football.  Running and being powerful and physical absolutely matters.  A lot.  Chip Kelly is great for the sport, but the Seahawks and 49ers were not slinging the ball around in shotgun last January.

It was ground and pound.  

Don't tell me for a second that they only run because they were ahead.  And that the old timers don't understand how football works these days.

Talk to Pete Carroll or Jim Harbaugh or the millions who watch what "works in today's game".  This, of course, should not be confused for Bob trying to bring back 80% run offenses.  Those days are gone and the great passing attacks are so good that they can replicate many great things about running with short horizontal passes.  But, that should not be mistaken for rendering the power run as antiquated or on the verge of extinction.

Now, back to my defense of Football Outsiders - if their actual point is directed at broadcasters, former players, and other people replacing tactical discussions with age-old cliches, then I am with them.  I think they are against people misapplying perceived truths rather than the overall summary that suggests running has very few positive impacts beyond nostalgia.

I recommend you read their stuff (as I do) and learn about their findings.  It is just on this particular one, I believe a lot of well meaning football enthusiasts are being misled about the actual effects of the running game and while it still owns a vitally important place at the table.


I realize that I called this post an email bag and then really didn't include much of an email (just a Twitter exchange up top).  So, here is one that I received that I think is worth reading about his findings as a statistician.  I think you will find that Richie and I agree on quite a bit.

Hello Bob,

My name is Richie and I enjoy your work on the Cowboys. I am also a lifelong fan of advanced metrics in sports. I work as a statistician for a living. In fact, I also provide a 'moneyball approach' to the game of golf for several PGA Tour players such as Ben Crane.

I was reading your recent article on how the statisticians are more or less, anti-running game. As a statistician, I used to interpret football in the same fashion. However, over the past 3 years I have started to see the folly in that line of thinking and how it is actually backed up by statistics. I'll try to give some bullet points and I think you will see how this applies to the Cowboys.

1. Historically defensive players have been more than twice as likely to get injured as offensive players. That is just a statistical fact. So, if you're throwing the ball around a lot, you are more likely to burn out your defense. Even if you're scoring tons of points, your defense is on the field longer. In order to stay healthy, the offense needs to be on the field longer because they are much less likely to get injured than the defense.

Forget about Dallas the past 3 seasons, if you look at the Patriots since they ramped up the passing in 2007, their injuries on defense have climbed dramatically.

That's a big reason why teams like Seattle and San Francisco make it to the NFC Championship...their defenses can stay healthier come January while teams like New England have depleted defenses.

2. While not a real statistical tidbit, the wind plays a large role in pass efficiency. So, if you're playing in December at stadiums that tend to get windy like Giants Stadium, the Linc or FedEx, you may have no offense if you are unwilling to run the ball.

3. Overall, passing the ball is more effective than running the ball. But, the standard deviation in results is much bigger. While there is more of a likelihood of gaining at least 8 yards on a passing play, the negative plays are entirely higher like a strip fumble, a sack, an interception or a holding call. Even an incomplete pass is a substantially negative play.

With running the ball, most carries are for 1, 2 or 3 yards. But, the odds of a RB fumbling are far less and the standard deviation in results is smaller.

It's kinda like comparing Barry Sanders vs. Emmitt Smith. Barry could certainly hit the home run and make the big play more often. But, Barry had the most negative carries in the history of the game. So the Lions offense would often start 2nd and 13. Whereas Emmitt may start the Cowboys at 2nd and 7. You need to use the run to help narrow the deviation in results so the offense doesn't have to make up huge amounts of yard in order to keep the chains moving.

4. The one metric I'm fascinated by is the one that shows the Cowboys have never lost when Murray gets 20+ carries in a game. We've seen similar incredible results for coaches like Joe Gibbs, Bill Cowher, Bill Parcells, etc. when they are able to run the ball 20+ times a game.

The common rebuttal to this is that it is a causation flaw. Obviously, you just can't run the ball 20 times in a row and expect to win.

However, the key metrics (carries and win percentage) are both extremes.

What I mean is...20 carries a game is not a lot of carries. Conversely, the win records of coaches like Gibbs, Cowher, Parcells and Garrett is nearly 100%.

20 carries a game means 5 carries per quarter. And if your team does run the ball 20 times in a game...that means they will likely have thrown the ball about 40 times. It's still a 2:1 pass to run ratio.

In order to get those 20 carries, the team can't fall behind by too much early on and the offense has to move the ball to some degree.

But, outside of those situations, the 20+ carries a game stat basically is showing the power of not neglecting the running game. Even if you're scoring points early on by passing the ball, it is critical to stick to the run. It narrows the deviation in results and takes time off the clock and keeps your defense fresh and wears down the opposing team's defense.

As much as I love statistics and I respect many sports statisticians, I think they have neglected these key points. Perhaps they are trying to drive home a certain narrative or they just haven't explored this enough.



Jay Baker said...

Great post, fascinating topic.
I also agree a lot with Richie there at the end.
I am extremely curious about the extent to which football teams use statistics and simulation ... *during the game*. That is, do they have sims running and stat points in place to help check themselves while the game evolves. If not, why not?

Doctorjorts said...

Bob, I think your response to the first email is destroying a straw man. FO doesn't say that running the football is useless. They are saying that NFL teams probably run too often when you look at the gains teams make from running plays vs. passing plays.

It all comes down to a crucial idea in the discipline of game theory called a "Nash Equilibrium." It's actually named for the guy Russell Crowe played in Beautiful Mind. Even if mixing your strategies is important, you should always favor the more successful strategy, and you should favor it more when it is more effective. The defense picks up on this and favors stopping your more successful strategy, until both are equally effective given a certain ratio of mixing. The fact that gains are still higher for passing than for running means that NFL teams still ought to pass more than they currently do.

Advanced Football Analytics has gone into great deal on the same topic.