Fun with Numbers: Why Running is Not the Answer

Special Featured Column By Bucs Gab Contributor Chris C.

I. Introduction, or Stats Are (sort of) for Losers

There’s been quite a buzz lately among Bucs fans in our attempts to pinpoint what’s going wrong with our team, and particularly with our offense, and even more particularly with Josh Freeman.  Just as there is a divide between those who would place the onus on Number Five and those who would place it on “Ollie,” there is a divide between those who see the Bucs as principally a running or a passing team.

Decidedly, the bulk of the commentary falls on the side of “run team.”  I am going to disagree with this, and I’ll try to do so without merely venting my spleen.  I think there is a strong case to be made for a different view, so even if the reader does not wind up agreeing with me, I hope at least to provoke some reflection.  This is going to be a long fortnight (sorry, still stuck in London mode) for us fans, so let’s take the opportunity to sit back, grab a cold one, and venture into the heart of darkness.

Maybe grab a sixer; this ain’t going to be pretty.

We should begin by admitting that the trouble with a lot of football commentary is that it is too absolute – “Do this, and voila! Success,” or “Don’t do this – kaboom!  Failure.”  Everyone knows deep down that the situation is more nuanced, and at a certain point we reach the intangibles that cannot be quantified, leaving us without a putatively objective way to put our arguments to the test.  I believe Raheem is largely correct when he says stats are for losers, but not completely.  Good turnover ratios are not for losers, for instance.  Yet even when stats are pertinent as indicators, they are so just because they are pointing to something that cannot be quantified – insufficient talent, bad coaching, low morale, lack of discipline, absence of “esprit de corps,” and so forth.  Those elements are ultimately what we are after when we use numbers to make arguments; they certainly cannot be taken at face value.

That being said, I think several commentators have been misled by how neatly our run yardage-to-victory ratio works out.  The correlation is real enough, but as any person who has been through the torture-chamber of Statistics 101 will tell you, correlation is not causation.  Keep that in mind as we try and make some back-of-the-envelope calculations in order to dig into the “internals” of the run-game data.

II.  What To Make of Our Run Game

The first thing to note is something that the diligent Sander over at Bucsnation recently noted: the correlation between run yardage, QB success, and victory faces a chicken-and-egg problem; namely, it doesn’t control for the fact that when teams are winning, they run more and get more yardage.  I would add that it also doesn’t take account of the ratio of run attempts to yardage – an important statistic in itself.  With a view to that, I did some digging and broke the numbers down as follows:

In our six losses in 2010 we averaged 18 carries per game and 78.3 yards; in our ten wins, average attempts were 24 and average yards were 108.1.  Nothing to scoff at; the disparity is real.

Thus far in 2011, in our three losses we average 13.7 carries per game and 44.3 yards per game; in our four wins those numbers shoot up to 22 and 113 per game, respectively.  Again, nothing to scoff at.

These are meaningful differences.  When averaged over the past 23 games, our losses saw 16.7 attempts and 67 yards per game, while our wins saw 23.4 attempts produce 109.5 yards per game.  In 2010 the average difference in carries between wins and losses was 6; in 2011, thus far, it is 8 – altogether the difference is about 7.  So the correlation is robust even when taking account of carries per game.

Why then do I refuse to accede to the notion that we are a run-first team?  Simple: these numbers do not license us to infer that.  What they do license us to infer is that a decent run attack is important – which everyone knew already.  The whole question, however, is whether we are a team that ought to be using the pass to set up the run, or using the run to set up the pass.  My answer is “both,” and which one we choose depends on who we are playing.

The intangible that this touches on is game-prep.  The many people I have seen make the point that our offense has grown too predictable are, in my view, correct, and this goes back to my analysis in my previous post.  The central strength of Raheem as a coach is his willingness to adapt; this is the element that was sorely lacking in Gruden’s system, and Raheem’s evident flexibility, even in the 3-13 season, was the main reason why I favored keeping him.  Olson’s offense, moreover, seemed to be moving in more creative directions every week last season, and it closed with a bang against Seattle and New Orleans.  This, I am saying, is because both Rah and Ollie were adapting the raw talent on their roster to the specifics of each upcoming team.  I think the numbers tend to bear this out.

Take a peek at the internals of our run stats.  In this season and last, we have had 9 games where we rushed for over 100 yards: Seattle (185), Detroit (148), Atlanta (111), San Francisco (140), Cincinnati (101), Carolina (153), Arizona (132), Indianapolis (151), and New Orleans (132).  We won 7/9 of those games, but that is not the only interesting factoid.  The other number to notice is the average ranking of these defenses (a difficult number to get exactly right, but the variance between methods isn’t all that large): 25th.  It would be difficult to scrape the barrel-bottom harder.

In 2010, Seattle (24th), Detroit (22nd), Cincinnati (25th), Arizona (23rd), and Atlanta (26th) were all bottom ten defenses, with an average rank of 24th.  This year, Indianapolis ranks dead last at number 32, while New Orleans molders at 25th.  The only by-the-numbers, bona fide defenses we steamrolled with the run were San Francisco (12th) and Carolina (9th), and factoring them into the 2010 data gives us an average of 20th for the ranking of our opponents’ defenses against which we ran for 100+ yards.

These rankings are based on a methodology used by John Stephens at Behind The Steel Curtain, and its main difference from the normal method is that Stephens takes account of yards per play (i.e., he controls for the amount of time a defense spends on the field – the more they’re on, the more total yards they’ll give up, making them look worse than they might be).

When using the more standard yards-per-game metric, the rankings don’t change much.  The 2010 overall (i.e., the defense as a whole) average-rank of our opponents who gave up 100+ yards in rushing to our offense stays exactly the same at 20th, while the average run defense ranking “goes up” (i.e., gets worse) from 17th to 19th.  When we include the 2011 games against Indianapolis and New Orleans, the overall average “falls” from 25th to 22nd, while the run average “rises” from 19th to 21st.

To sum this up, the average overall rank of defenses we ran 100+ yards against in 2010 is about 20th, while the average run-defense ranking of the same set of teams falls between 17th and 19th.  This year the sample size is only two games, but their average overall defensive ranking is still revealing: 29th.  Their average run defense ranking is 28th.

Don’t be put off by all of the numerology – there’s nothing complicated about what I’m doing.  I’m simply calculating averages on numbers people don’t usually take into consideration.  Add ‘em up, divide by n.  Anyone can do it.  The upshot is what is important, and that is evident to the naked eye: we ran pretty well against pretty bad defenses.  But how good were the defenses we ran poorly against?

Let’s investigate.  Using Stephens’ numbers, the average overall ranking for defenses we ran under 100 yards against in 2010 is 16th; for the run defenses the average is 15th.  In 2011, so far, the numbers are 10th and 8th, respectively.  Combining the two seasons, we get an overall average rank of 14th and a run-defense average of 13th.

All we do now is compare and contrast.

100+ Rushing Yards (average rank of opposing defenses)

2010: Overall: 20th; Run: ~17th and 19th

2011: Overall: 29th; Run: 28th

Combined: Overall: ~22nd and 25th; Run: ~19th and 21st

Less Than 100 Rushing Yards

2010: Overall: 16th; Run: 15th

2011: Overall: 10th; Run: 8th

Combined: Overall: 14th; Run: 13th.

Of course, this is football, and as in the real world there are outliers all over the place (which is why I broke the 2010 data up from the 2011 data), but these cut in both directions.  One could also criticize me for using 100 yards as the standard for “running well,” but the data would look even more skewed if I picked a number like 80 or 60.  For instance, if we choose 60+ yards as our standard for a “good” running game, the overall and run defense averages of those teams who held us under that number in 2010 are 9th and 11th, respectively.  In 2011, it is 9th and 10th.  With tons of games skewing the 60+ yard side of the table, all this would do is confirm my point by a statistical trick.  Drawing the line at 100 yards is thus an attempt to make the data speak something like the truth.  I think the result of the exercise, as presented above, tallies with what we would have expected based on what we’ve seen.

Furthermore, there is still a difference between run attempts in our above- and below-100 yard rushing games: above, the average is 24; below, 19.  Not a huge difference, but it surely matters to some degree.  And to reiterate, no one is denying that a good running game is important.  The question is how our offense should be oriented.

Return to the numbers and take stock for a moment.  One data point is that in the past 23 games, we are 7-2 when we run over 100 yards.  There is also an average difference of 6 carries and 43 yards in rushing between our wins and losses.  Further, we do carry it more – by an average of +5 – when we run over 100 yards.  All of this points to the importance of the ground game.

On the other hand, since 2010 the difference in quality between the defenses against which we’ve run for 100+ yards and those we haven’t is considerable.  Put simply, if your defense ranks 20th or worse, we will get our 100 yards; if it ranks 14th or better, we will fold like origami.

What are we to make of these two data sets, one indicating we should run and the other indicating we are in fact rather bad at it?  Ask the question in another way: If we are facing a top 15 defense, should we pound the run?

This is the rub.  As I remarked earlier, what I like about Raheem and what I (used to) like about Olson is their apparent flexibility, their complete devotion to the philosophy of adaptation.  They have at least seemed to understand that whatever pre-game advantage our team has will be based on the element of surprise – also known as unpredictability.  The 49ers used the same tactic against us earlier this season, to devastating effect.  Perhaps what subconsciously appalled us so much about that game, aside from the outcome, was the horrid experience of seeing another team do what we were supposed to be doing.  Likewise, perhaps we were so exhilarated by the subsequent Saints game because we seemed to come back with the adaptability and capacity to surprise that is the strength of this team.

All hypothetical, admittedly.  What is not hypothetical, however, is that we run good against the bad and bad against the good.  It is by no means clear that we can simply defer to conventional wisdom and assert that we just need to pound the rock harder against the good, and that this will lead Freeman to improve.  It’s a real “Be careful what you ask for” situation.  For not only do we not know if our running backs are capable of doing this, we also do not know if the main reason we didn’t run against good defenses was because we tried and failed, and then fell behind.  My recollection, in any event, is that that is more or less what happened.

One point on which I agree with those who want to see more running is that we are not committed to it.  What I mean is that, as I’ve put it before, we are getting the worst of both worlds.  If Olson (or whoever) wants this to be a slow-moving, pound-it-home offense, then he shouldn’t be abandoning the run early.  That’s tantamount to abandoning order for chaos.  The offense isn’t designed for chaos right now, so when we keep the pound-the-rock conventional approach but tell Freeman to go out there and sling the guns, we’re asking him to, in effect, work unconventionally within a highly conventional scheme.

What we have now is gunslinging without the play calling or dynamism necessary for it to strike bulls-eye.  That’s what predictability is: Freeman will drop back, stay in the pocket, and fire away into the haze of the Cover Two.  So, this much is true: only a solid run game can suck in a defense enough to make success in that style feasible.  Look at what happened to Matt Ryan when we shut down Michael Turner and got a pass rush going.  It’s exactly what’s happening to Freeman almost every game.  And that’s because, for God knows what reason, we are trying to become Atlanta-esque on offense.

I’d remind my readers here of what was so appalling about the Gruden Bucs near the end: They were being used as though they had the best running back and the best o-line in the league.  Instead, they had Cadillac Williams, Earnest Graham, and a middling line.  The result was some impressive games against sub-par defenses, and some horrific showings against above average ones.

I suggest that the same applies today.  Blount is good, but he is not the kind of back who can sustain a full-on run attack against any and all comers.  Our o-line is decent, but the same caveat applies.  Moreover, Freeman is quite evidently not comfortable in that kind of a system.  His receivers aren’t either.  “Run the ball” is not the answer – it will take Raheem and his team straight into late Gruden territory.  The answer, in my view, is easy to state but hard to implement: vary the focus of the offense based on who we’re playing.  We can presume the bread-and-butter style will be relatively successful against the sub-par.  But we can also presume it will not work against the above average, and definitely will not work against the truly good.

I would cite the 49ers game as the textbook case-in-point.  We disregarded the data I cited above, probably because we were feeling hubristic given that we ran over the Niners last season.  Bad move.  We came out and tried to pound the rock, predictably got stuffed, fell swiftly behind, and then had no bloody clue what to do.  Furthermore, the Niners knew we were going for a more conventional approach this season, and we knew that they knew.  It was just plain arrogant to think that we could go out there without tweaking our strategy in any way and then steamroll them, because we’re just that good.  For those who want us to double down on that M.O., all I can say beyond what I’ve already said is: We are just not that good.  And in sport, where talent does not tread brains must rush in.  We need far more brains, far less presumption of talent – or rather, far more brains in the use of our talent.

III.  Objections and Rejoinders

One obvious retort to my analysis so far is that while we may not have run well against above-average defenses, we also didn’t pass well.  This is true to an extent, especially against the very best (Pittsburgh and Baltimore in 2010; San Francisco this season).  If we recall the Pittsburgh game, we fell behind early; we also did not have Blount.  Cadillac and EG could not run on that defense, no matter how much we tried to will it to happen.  Against Baltimore, Freeman was off in a game that was winnable.  It was cold and everyone looked out of their element until the fourth quarter.

Still, we ran for 53 yards against Pittsburgh on 14 carries in a 38-13 loss; by way of comparison, we ran for 52 yards on 30 carries in a 20-7 victory over Carolina (remember that the Panthers had a good run defense in 2010, ranked 9th, and we did manage to steamroll them on the ground after Blount arrived).  Against the Ravens we had a not-dreadful 74 yards on 17 carries in a 17-10 loss, which numbers were good enough to help us beat New Orleans 23-13 later in the year (73 yards, 20 carries).  Note as well that we ran it 20 times for 77 yards against San Francisco this year.  Against Minnesota, we ran 16 times for 89 yards – not attempting more than against the Niners, just running better.  A similar game to the Minnesota victory this year was the St. Louis match-up last year – 16 attempts, 84 yards, a slim victory pulled out at the end.  The reader will notice that there is no pattern here except the obvious one: We comprehensively struggle on offense against good defenses.

We now come back to the core issue – should we be trying to go Atlanta-esque in order to clear the hurdle against good defenses, or should we be doing something more idiosyncratic and tailored to the talent we have?  My answer is the latter, and the reason is that Freeman is our linchpin.  If we mean what we say in delegating the franchise to him, then we need to create the infrastructure for him to do what Brees, Brady, Rodgers, and Roethlisberger do, and in his own way.  I can’t stress enough that I don’t deny that the run is important; what I assert is that to make it our centerpiece, the core of our identity as an offense, the sine qua non of our success, is not the way to go – unless, of course, we want mediocrity as far as the eye can see.

The run-centric approach amounts to saying, “We’ll bet on Freeman, so long as everyone else steps up first.”  This statement should, I believe, be flipped around: “We’ll bet on everyone else, because we’ve got a system that allows Freeman to step up.”  If he’s our commander, let him command.  Notice that this is a problem in general with our staff’s handling of our offense – Olson is scared by Blount’s inexperience, so he denies him the experience he’ll need to gain confidence.  He doesn’t trust Benn to do what he should on passing plays, so he yanks him.  It would be no shock to find that he doesn’t trust Freeman to take command of a fluid, adaptive offense, and as a result hog-ties him to the pocket and force-feeds him play calls.  If this is what is happening, then Freeman is not being trained to be a commander.  That has to stop.  Even if sticking Blount in for checkdowns and running Benn on pass plays and putting more discretion in Freeman’s charge results in mistakes and some sloppy play, we have already seen the alternative: mistakes and sloppy play.  What exactly is there to lose by letting our very best – Freeman, Blount, and Benn – go out and play football?  We might stink up the place?  We already do!

Franchise players are the heart of their teams.  If what we actually want to do is create a system with no core player, no person who has the reins, then the approach of not going after the best free agents is self-defeating.  To be a team that uses Freeman as a compliment, a sort of cherry on the sundae, we obviously need a sundae for him to top off.  Clearly, our management has no intention of getting that; hence, asking the team to simply play like delectables despite not having the ingredients for it is only going to produce a bitter stew that Freeman, at best, adds a bit of salt to.  Olson seems to want pop rocks; Freeman, Benn, and others want to go nuclear.  Why not let them try?

Along these lines, another claim I often see is that our offense passes too much – and it is true that Freeman has more attempts this season than anyone outside of Drew Brees.  However, that is more of a testament to how unusual this season has been than to anything anomalous in our passing system.

Freeman played 10 games in 2009 and threw 290 passes.  He’s played 7 this season and thrown 270 times.  In 2010, he had 474 attempts.  On a linear projection, indeed, this season looks anomalous: Freeman is on pace to throw 617 passes, with an average of 38.57 per game.  But hold on.  If we’re going to cite statistics then we need to use them properly, and these numbers are mostly artifacts of outliers – there was nothing to do against Chicago but throw 51 times, and the Detroit game (43 throws) was a mistake acknowledged by everyone.  The only other game Freeman threw over 40 times was against New Orleans (41), and I didn’t hear anyone complaining about that (the run attack was also strong).

What this picture needs is perspective.  In the first 7 games of 2010, Freeman threw over 40 times twice (compared to three times this season); over 30 times twice; and under 30 three times.  It’s important to remember that those under-30 games saw our passing game at its best.  That is, Freeman threw it less because he hit receivers for big touchdowns, thus eliminating the need for prolonged drives.  As few as one of those in a game can cut your pass attempts by 6 or more.  Against Arizona and Carolina, there was more than one big play.  The safe inference, then, is not that Freeman is, all things considered, throwing too much this season – rather, it is that our passing game is not performing up to snuff.  When you bracket the Detroit game (which was played as though Blount was injured) and the Chicago game (where we had no capable running backs), and then take account of the inefficiency of our passing attack in general, Freeman’s passing attempts are not out of sync with last year.  In games with Blount at full health and integrated into the offense, Freeman’s attempts are 31 (Minnesota), 32 (Atlanta), and 39 (Indianapolis).  With just one or two drives such as those we had early against Cleveland, Carolina, and Arizona last season, those first two numbers would probably have been in the 20’s and the last would have been in the low 30’s.  The problem is deeper than “passing too much,” a charge that is simply the flipside of the notion that we are not running enough.  The question here is: When we pass too much, why do we do so?  Sometimes, as in the Detroit game, we can all agree that the cause really was not enough running.  At other times, as in the Indianapolis game, the cause was clearly poor performance (we attempted 27 runs that game, so we definitely did try to run the ball).  At still other times, as against Chicago, there’s been an obvious need to pass the ball in addition to our futility problems.  Bottom line: the picture is complicated.  On balance, it seems that the root of the problem is an underperforming passing attack, not passing too much or avoiding the run per se.

IV.  The End (at last!)

Considerations like these are why it is important not to be misled by the apparent story bare numbers tell on their face.  What we want to know is what our eyes see: Why does our passing attack stink?  If Mike Williams doesn’t get that touchdown called back earlier in the season, and Benn doesn’t step out of bounds against Indianapolis, and Williams catches that touchdown against Chicago, things would look a bit different right now.  At a more fundamental level, we cannot forget that there is another possible explanation for our failures on offense aside from too much passing/too little running; namely, the possibility that our offensive mold has been cast in a way that does not fit the players we have.  I think this is probably the right explanation.  And it means that in order to fix what ails us we will need far more basic changes at the conceptual and playbook level than merely concocting a different mixture of runs and passes.

In conclusion: against the poor-to-mediocre, we can try orthodoxy; but against the above average-to-excellent, we need heresy – put Freeman in motion, use the pass to open up the run, and be surprising with the routes and the players running them.  Against the lesser we can win with brawn; against the mighty we must win with brains.  At this point, given our poverty in the backfield, we don’t have much of a choice anyway.

Bucs Central: Thanks to Erik Sabol for dishing Bucs Gab some major kudos on his site!

Enjoyed this post?
Subscribe to NFL Gridiron Gab via RSS Feed or E-mail and receive daily news updates from us!

Submit to Digg  Stumble This Story  Share on Twitter  Post on Facebook  Post on MySpace  Add to  Bark It Up  Submit to Reddit  Fave on Technorati

Comments are closed.