Why Passing Beats Running

[BobDigital]

Passing is better than running clearly. Just look at the average pass play vs the average run play and you will see the average pass play on the whole leads to a better out come particularly when we fact in natural abbe ration between plays.

The average YPP is something like 7.2 and the average YPR is 4.2

So about a 3 yard difference per play between running and passing. Pretty clear which is the overall better option most plays.

 

[maineman209]

I just ran the league numbers for 2016.
... League average pass percentage was 58%.
... Highest was BAL at 65%, followed by DET and NOL at 63%.
... Lowest were BUF and DAL at 49%, followed by TEN at 51%.
... The 2016 Pats were at 53%, 5th lowest pass percentage in the league.

The 2015 Pats (with 17 RBs and 35 OL on IR) were at 62%. The last time they were above 58% was 2002 with 61%.

The 2007 season was the first in which teams were allowed pre-game possession of game balls so that QBs could "customize" (rough up) the surface to their own liking. Prior to 2007, officials took brand new game balls right out of the box - wax coating and all - just before kickoff.

It was also in 2007 that the league average passing percentage began to climb.

In 2007 thru 2014, the Pats passing percentage has been 57%-58% about half the time and 55% or below about half the time.

Prior to 2007, the Pats passing percentage averaged 53%, even including the anomalous 2002 season of 61%.

Pats Superbowl seasons:
2001 = 50%, WON
2003 = 53%, WON
2004 = 48%, WON
2007 = 57%, lost
2011 = 58%, lost
2014 = 58%, WON
2016 = 53%, WON

 

[Deus Irae]

If you read the study and chew on it a bit, you'll see that all of the situational factors mentioned above are taken into consideration. The authors are as surprised as anyone else at the outcome.

What's surprising is that people still take this stuff seriously, when it's about 1% legit and 99% nonsense.

 

[Deus Irae]

I think this is an issue of abusing statistics.

Consider the title "Average Win Percentage added for rushing and passing plays on 1st and 2nd downs with 5-10 yards to go."

(1) Pretending 3rd down doesn't exist seems rather problematic to me.
(2) Also lumps for example 2-5 and 2-10 together. It seems pretty clear to me even without statistical analysis that running on 2-10 is much less likely to lead to a good outcome than running on 2-5.
(3) In some sense this is a pre-determined result. Any given pass play is likely to have a bigger result than any given running play, and therefore likely to have a bigger WPA I would assume.

Then looking at the pretty graph I think the bubbles are misleading. Consider the 2 large bubbles for passing and running for the 1st quarter tied score. According to the bubbles no play will ever produce a negative WPA. This is clearly false and a result of statistical massaging. Which really makes it impossible to draw any conclusion from the graph.

It's very similar to the way the "Go on 4th down!" data was abused.

 

[maineman209]

Passing is better than running clearly. Just look at the average pass play vs the average run play and you will see the average pass play on the whole leads to a better out come particularly when we fact in natural abbe ration between plays.

The average YPP is something like 7.2 and the average YPR is 4.2

So about a 3 yard difference per play between running and passing. Pretty clear which is the overall better option most plays.

OTOH, league average is one turnover every 44 pass attempts and one turnover every 139 rushing attempts. Risk/reward.

Of course, with the 2016 Pats, it was one turnover every 275 pass attempts and one turnover every 241 rushing attempts (and Belichick is probably steamed about the latter).

 

[SlowGettingUp]

minimizing potential negative consequences of a play

The trouble is that for many coaches, the main negative consequence they worry about is the chance of getting fired rather than the chance of winning the game.

If you are a significant underdog, your best bet is to increase the variance of the outcome. So go for more risky plays.

The place where you should see this, but don't, is in the 2-point conversion. Recent statistics have shown this to be pretty much a wash - your point expectation is about the same as trying the kick. So underdogs should pretty much always be going for 2, even early in the game - but they just don't.

The other high-variance plays are onside kicks and going for it on fourth down. Again we don't see nearly as much of these from weak teams as we should see - and that is because their coaches are either dumb or scared of getting fired.

 

[QuantumMechanic]

Sorry if we're getting too far afield here, but this is what excessive risk aversion is all about: how heavily you weight minimizing potential negative consequences of a play, as opposed to maximizing chances of winning the game. Those two factors sometimes point toward different decisions.

Exactly that!

Of course, that's because a coach's primary goal (even if he won't admit it to himself) is to keep his job, not to win games. (Though obviously the latter helps the former!)

Between the mediots and risk/statistically-illiterate fans (and probably owners) a coach will get roasted alive for making a decision that maximizes the chance of winning but does not minimize the chance of a short-term bad outcome and that decision doesn't pan out.

This happens all the time in end-game scenarios where coaches prioritize postponing the moment of certain loss over maximizing the chance of winning the game because it's better PR to do the former.

 

[BobDigital]

OTOH, league average is one turnover every 44 pass attempts and one turnover every 139 rushing attempts. Risk/reward.

Of course, with the 2016 Pats, it was one turnover every 275 pass attempts and one turnover every 241 rushing attempts (and Belichick is probably steamed about the latter).

Even with everything factored in passing still beats running as a means of getting yardage and scoring.


Yes if you pass turnovers are more likely to happen but it is a reward that far exceeds the risk particularly when you need to make at least 10 yards every 3 plays.

 

[Nehalem]

Passing is better than running clearly. Just look at the average pass play vs the average run play and you will see the average pass play on the whole leads to a better out come particularly when we fact in natural abbe ration between plays.

The average YPP is something like 7.2 and the average YPR is 4.2

So about a 3 yard difference per play between running and passing. Pretty clear which is the overall better option most plays.

The problem is you win football games by scoring the most points; not by get the most yards per play.

Which is the real problem with the article. In football more so than other sports the result of previous plays impacts the current play. So that an action that might be suboptimal over a 1 play span, can actually be optimal over a multiplay span.

 

[patchick]

The trouble is that for many coaches, the main negative consequence they worry about is the chance of getting fired rather than the chance of winning the game.

Between the mediots and risk/statistically-illiterate fans (and probably owners) a coach will get roasted alive for making a decision that maximizes the chance of winning but does not minimize the chance of a short-term bad outcome and that decision doesn't pan out.

This happens all the time in end-game scenarios where coaches prioritize postponing the moment of certain loss over maximizing the chance of winning the game because it's better PR to do the former.

Absolutely. In the classic XLIX example, the Seahawks' coaches were demonized largely because they made the less expected move of passing at the goal line. Whereas if they'd asked Lynch to run straight into a d-line stacked against the run and he'd fumbled it away, I'll bet the coaches would have been held blameless and the players demonized instead. You seldom get blamed for making the conventional call, whether it's the smart call or not.

And if Seattle hadn't just gone to back-to-back Superbowls -- if, say, that game dropped their team to 8-8 and just out of playoff contention -- the aggressive play call might have cost the coaching staff their jobs.

So I guess I should add that to my list of reasons you might want to call a running play: job security.

 

[BobDigital]

The problem is you win football games by scoring the most points; not by get the most yards per play.

Which is the real problem with the article. In football more so than other sports the result of previous plays impacts the current play. So that an action that might be suboptimal over a 1 play span, can actually be optimal over a multiplay span.

Well of course you generally need to run somewhat to keep balance but most scoring drives don't end up as scores cause of running plays but passing plays.

IMO in a perfectly run offense. You run the ball to keep the defense from going too light and the pass rushers honest as well as to set up play action. You do it to get short yardage gains (2nd/3rd and short) and hopefully accomplish all that while running effectively. But those runs don't tend to really move you down the field much. You move down the field generally with pass plays.

 

[Deus Irae]

Well of course you generally need to run somewhat to keep balance but most scoring drives don't end up as scores cause of running plays but passing plays.

Most scoring drives end up as scores because of a combination of running and passing plays.

 

[ivanvamp]

This is an interesting study. But it's based on the reality that teams run the football a LOT. Not most of the time, but a LOT. If they threw 90% of the time, say, then what would happen is that teams would play smaller, faster units designed to cover and rush the passer, and then passing #s would go down, and running would become more effective. Then another study would come out and say that teams need to run more.

The NFL tends, like a free market system, to find the sweet spot.

 

[ivanvamp]

There's also the issue of turnovers. In the NFL this past season, there were 285 fumbles and 404 interceptions. That's 199 more INTs than fumbles. Moreover, those 285 fumbles includes QB strip sacks (which should really count against the passing game), fumbles by receivers after making a catch (which should also count against the passing game), and kick returns (which should be taken right out).

Long story short, football isn't all about moving the ball and scoring points. There are plenty of situations where all you need to do is take care of the ball.

Moreover, sacks are a big factor. A loss of yardage on a sack is much higher than the average loss from a running play.

Also, there's the clock. Sometimes you just want to keep the clock moving, and so passing brings with it an increased risk of stopping the clock.

Passing, when it works, has a much higher likelihood of moving the ball down the field, but it also comes with increased risk. The more you pass, the more likely it is that you'll have a massively negative play that could have a big impact on a game, not in your favor.

 

[SlowGettingUp]

There has generally been a trend to smaller, faster LBs and much more nickel/dime.

Let's see if Belichick now zigs when everyone else zags, and starts running the ball more. He certainly has paid up a lot more this season for RBs than he has in recent memory, and he should also have two TEs both highly capable of blocking. Add an OL with an additional year of experience and perhaps the desire to reduce wear-and-tear on Brady.

 

[Tony2046]

Well of course you generally need to run somewhat to keep balance but most scoring drives don't end up as scores cause of running plays but passing plays.

IMO in a perfectly run offense. You run the ball to keep the defense from going too light and the pass rushers honest as well as to set up play action. You do it to get short yardage gains (2nd/3rd and short) and hopefully accomplish all that while running effectively. But those runs don't tend to really move you down the field much. You move down the field generally with pass plays.

I disagree. Moving the ball and scoring consistently requires both. Game management requires both. Each play is situational and dictated by what the defense is giving you which can be determined by which defensive packages/personnel the defense has on the field and the mismatches they create respective to your offensive personnel. Teams are built differently which creates different mismatches each week depending on the opposing teams personnel e.g, Cowboys vs Saints.

Below are the top scoring offenses from the 2016 season. Rushing % ranges from 21.57% to 63.04%.

Which could bring us into a discussion of how the offense compliments the defense. For example, if the defense sucks will offense attempt to sustain longer drives to give the opposing offense less time on the field?

 

[Koma]

I just ran the league numbers for 2016.
... League average pass percentage was 58%.
... Highest was BAL at 65%, followed by DET and NOL at 63%.
... Lowest were BUF and DAL at 49%, followed by TEN at 51%.
... The 2016 Pats were at 53%, 5th lowest pass percentage in the league.

The 2015 Pats (with 17 RBs and 35 OL on IR) were at 62%. The last time they were above 58% was 2002 with 61%.

The 2007 season was the first in which teams were allowed pre-game possession of game balls so that QBs could "customize" (rough up) the surface to their own liking. Prior to 2007, officials took brand new game balls right out of the box - wax coating and all - just before kickoff.

It was also in 2007 that the league average passing percentage began to climb.

In 2007 thru 2014, the Pats passing percentage has been 57%-58% about half the time and 55% or below about half the time.

Prior to 2007, the Pats passing percentage averaged 53%, even including the anomalous 2002 season of 61%.

Pats Superbowl seasons:
2001 = 50%, WON
2003 = 53%, WON
2004 = 48%, WON
2007 = 57%, lost
2011 = 58%, lost
2014 = 58%, WON
2016 = 53%, WON

One note: The rule allowing QBs to prepare their own footballs went into effect for the 2006 season.

Here's an article by Peter King from September 2006 (Be prepared, the formatting is terrible, lots of words running together) about how Brady and Manning spearheaded the effort in the offseason after the '05 season and the new guidelines being approved in March of '06.

 

[Ring 6]

Sorry if we're getting too far afield here, but this is what excessive risk aversion is all about: how heavily you weight minimizing potential negative consequences of a play, as opposed to maximizing chances of winning the game. Those two factors sometimes point toward different decisions.

But you characterized them as too conservative for disagreeing with a statistical tally that doesn't take that into effect.

 

[RayClay]

So did you just not read the actual words under the fancy chart or did you outright ignore them? The OP says there are situations where you have to run and outright specifies that.

I don't think he made a relevant point. Of course you gain more yards passing, that's why they don't lateral like in the leather helmet days. Not really worth spending the effort he did to make an obvious point, especially since his statistics don't really prove anything. If a team only ran when he said it was advantageous, that would negatively affect passing plays because defenses would adjust. There's a difference between compiling statistics and proving a point. Also, it greatly depends on who your QB and RB are.

 

[RayClay]

2001 Patriots are a great example of using marginally successful runs to set up the run later and keep pressure off QB.