Identity Crisis

If there’s one meaningless word thrown around by football analysts in the last couple years, it has to be identity. “The Jets offense has an indentity crisis.” “The Ravens need to find their identity on offense.” “The Eagles have lost their offensive identity.”

Are we talking about professional football teams or teenagers trying to figure out whether to hang out with the jocks, dweebs, preppies, or wastoids?

Identity has replaced rhythm as the most meaningless word in the NFL. Remember those days when every broadcaster at one point in the game had to say, “The 49ers need a couple of completions here to get into a rhythm"?

What the hell does that even mean?

I’ll tell you. Identity, rhythm, and other nonsensical words are just silly replacements for the word good. The Packers offense is good, therefore it has an identity. Against the Vikings last week, “they found their rhythm.” The Jets offense isn’t good, so they don’t know if they “should have a passing or running identity.” Against the Broncos last night, they “never could find their rhythm.”

On TV, you just can’t simply say a team is good or not good and leave it at that, but that’s all these masters of the obvious really have to say. So football pundits dress up their banal observations with voodoo nonsense so we viewers think they have some sort of deep understanding about the game that we don’t.

Everyone understands that timing is critical in the modern passing game, and I suppose that’s what people could mean when they refer to ‘rhythm’. But they don’t. They don’t explain how a receiver might be coming out of his break a half step too early or how a quarterback might be rushing his drop back. And when they talk about identity, there’s no accompanying analysis about what opposing defenses have allowed, or whether the running game is really any good to begin with. We only get a quick reference to how few carries [overpaid running back] got, and how poorly [unprotected quarterback] did in the most recent game, which was a loss. End of analysis.

I could write a computer program in five minutes to do a better job of identity analysis than any sports page in the country. In fact, I just did.

The table below lists the relative performance of running and passing for each offense. Performance is rated by Expected Points Added per Play (EPA/P). Also listed is each offense’s percentage of pass plays. Building off Carson's earlier idea, all data is limited to plays that occur when the offense’s win probability (WP) is between 0.15 and 0.85. This excludes situations in which a team is either well ahead or behind, forcing a pass- or run-heavy strategy mix.

On the right is each offense's split between passing and running performance. The higher the split, the better the team has been at passing. Negative splits indicate the team has been better off running.

TeamPass %Pass EPA/PRun EPA/PTot EPASplit

Normally, game theory considerations would suggest that to optimize overall production, teams at the bottom of the list should tend to run more often and teams at the top of the list should pass more often. In a pure point-optimization sense, that's true. But lesser teams need high variance to beat better teams, and therefore need to pass. It's a paradox that the worse an offense is at passing, the more it needs to do it.

But we can be relatively certain saying there's not much reason for teams near the top of the list to lean more heavily toward running the ball. So before you pen your column about the glorious wonders of smashmouth running games and imposing your will and all those other trite football sentiments, consult this table.

There are many factors--defensive schemes and looks, personnel packages, injuries, play action considerations--that dictate run-pass balance. But if, on average, every run is losing your team points and every pass is gaining your team points, calling for more running makes little sense.

  • Spread The Love
  • Digg This Post
  • Tweet This Post
  • Stumble This Post
  • Submit This Post To Delicious
  • Submit This Post To Reddit
  • Submit This Post To Mixx

23 Responses to “Identity Crisis”

  1. Anonymous says:

    Very interesting article. Nice job. The disdain for Football Talking Heads is shared. Another absolutely empty qualifier seems to be 'elite' which has gone from being exclusively applied to the QB to just about every position and unit on the field now.

  2. Anonymous says:

    Fascinating article. Any chance that this chart will be updated later in the year or even possibly become a weekly feature? Great work. Jason

  3. Brian Burke says:

    The 'elite' thing is killing me!

  4. Sam P says:

    I think there actually could be a definition for identity. It's very unlikely for a team to be equally good at everything, and more likely to be good at a very limited range of things and mediocre or worse at everything else. That profile along with what you attempt to do is your identity. Hence we talk about a team running a single wing offense (though not in the NFL, heh) or a Tampa two defense and so forth.

    [now what happened to my Sam's Hideaway login...]

  5. Pete says:

    My favorite "what the hell does that even mean?" term: "physical" as in "a very physical player/team/rivalry. If it consistently meant the opposite of finesse, it would be fine, but announcers seem to use it as a lazy term that could include a team that runs the ball a lot, players who block, players who tackle, players who are big, teams that boast that say are physical, two teams that are good that play each other a lot and don't like each other--basically a blanket term describing just about everything that happens in football all the time that's not punting.

  6. Brad Warbiany says:

    Like Sam above, I'm not completely sold on the idea that "identity" is a meaningless term.

    I look at "identity" as what a team is good at. If they align their playcalling with what they're good at, it could be said that they've found their identity. If they don't align their playcalling with what they're good at (i.e. if they do something like try to use Tim Tebow as a pure drop-back passer), one can say they haven't found their identity.

    Coordinators who try to align their players to a pre-set "system" rather than utilizing the talents of their players in their most optimal fashion could be said to have an identity crisis.

    Essentially (and I follow Chris Brown @ smartfootball rather closely), an offense should have an identity. It should have it's bread & butter plays, that it's built around and that are expected to work against any "base" defense. And then it should then have its constraint plays, to exploit defenses that try to cheat to take away the bread & butter.

    This is certainly more of an issue in the college game, where the variability in player talent and skill set is much higher, of course. One can't say that Texas Tech under Mike Leach or that Georgia Tech under Paul Johnson don't have an "identity". But my preferred team [Purdue] is one that I think attempts to do too many things at the expense of doing any of them well. They don't have an "identity" of what they do really well, and although most fans see some big promise in the running game, the coaching staff seems insistent on forcing an average QB to try to excel in a spread scheme. It's a team that tries to pass to set up the run, when the skill set of the team might dictate that they should run to set up the pass. That's what I would define a "lack of identity" as.

  7. stevekirsch says:

    I'm with Sam P here. I think that finding an identity is identifying what your team strengths are and leveraging those through an effective game plan. I think Brian's table does a good job of identifying team strengths. There's some additional depth to these (inside runs, outside runs, deep passes), but they highlight what works for a team. Most sportscasters would say a team looses their identity when they go away from their gameplan of what works for them. Whether or not this is beneficial or not is another question. The data seems to indicate that even if your identity is to run alot, it is better to pass when the underdog/behind.

    I'd also like to take a crack at rhythm here. I think this relates to playcalling/gameplanning as you are getting the defense to react how you want them to. Its about using the run to setup the pass and vice versa. Of course it all starts with successful plays and execution, but there is an element to keeping the defense on it heels by effectively mixing up the play calls. From a statistical perspective, we look at each play in a vacuum, but play action shows us that the success of a play can have additional effects and benefits on later plays. A team has to be good to develop "rhythm", but there is a another playcalling/coaching/gameplanning aspect that helps exploit strengths/weaknesses and adjust as the defense adjusts.

  8. Adam says:

    Great stuff. Drives me nuts every time they tell me the Giants need to go back to being a running team. Um, as anyone actually watching Eli pass vs. Jacobs plowing into the line for 2 yards would agree, no, they don't.

  9. bob says:

    slightly off topic, but a note on the tables. The rank gets reversed when you reverse a column, those should stay as they always are.

    for instance, if you look at qb hits, it shows the most qb hits, pitt #1. then click the column top to list it by least qb hits, pitt goes to the bottom (as it should) but it is still labeled #1.

  10. Jim Glass says:

    Finally, the first Brian Burke post I've really disagreed with in a good four years!

    Not that I disagree that commentators forever throw around mean-nothing terms like "rhythm". As a Jets fan, endless complaints that "Schott doesn't know how to call an O with any rhythm" drive me batty. (Is football "dancing with the jocks"? Did Lombardi have his O practice the Packer Sweep to a metronome, "c'mon Kramer, step to the beat."?)

    But "identity" has real meaning, IMHO. In Eastwood's words, "A man has to know his limitations." In statistical terms, a coach has to know what investments (in cap, roster spots, play calls) produce the highest marginal returns, which lower and which negative. The pattern that produces the highest marginal returns for *your* team given *its* personnel and abilities is "its identity". That's different for every team. Losing track of that is costly. And "The Jets offense has an indentity crisis,” is something I definitely agree with.

    Consider: Rex takes over, builds a very top D, finds himself with a kid QB who has good physical talent and good character/intent, but is a novice and terrible decision maker. When the game goes as prepared the QB works well off the game plan. But if he has to improvise and throw a lot to "carry" the O the roof collapses on him. (His A/YA-to-attemps correlation: -23% -- diminishing returns).

    Fitting that, Rex builds a big "Ground and Pound" running game with a top O-line and three very good RBs. With 172 yards and 148 yards a game rushing, Sanchez was able well enough within his efficency bounds that, added to the rushing, it gets them to the AFC finals twice. That is the O's "identity".

    Then Rex spends this last off-season chasing Nnamdi Asomugha (because with Revis and 2 other above-average CBs he needs another above all things. Talk about marginal return.) The Jets seriously weaken the OL to save money and go into the season with only 1 1/2 RBs -- one part-timer and a second-tier full timer. Rex *inexplicably* decides that's OK because, in spite of no evidence of any more than incremental improvement on Sanchez's part (and after getting rid of the two WRs he played with during his two years) it is time to go "Air and Dare".

    Predicitably that's a fiasco, so after four games Rex publicly decides it's back to the ground and pound. But now that's worth only 90 a game -- and heavy rushing with only 1 1/2 backs the result is yesterday's game: a special-teamer playing RB behind a sieve of a line, and an O that can't out-play a Tebow team.

    Because Rex took the O *far away* from what it did best, it's "identity", and straight into declining returns.

  11. Jim Glass says:

    After my first disagreement with our host, a note on how much I do agree in general about all the other meaningless words pundits throw around.

    I haven't seen this quoted here yet. It's from the Michael Lewis story about Daniel Kahneman, who won a Nobel for investigating how people make decisions, and Kahneman's partner Oren Tversky who died before he could share the prize. And how their work led to "Moneyball" thinking in baseball (which from there is now spreading to football). Lewis writes:

    Tversky’s son had been a student in a seminar I’d taught in the late 1990s at the University of California ... [he] put me in touch with his mother, Barbara, who put me onto the papers left behind by her husband...

    Then one afternoon I came upon a letter dated June 4, 1985, from Bill James. The baseball analyst whose work was then being blithely ignored by professional baseball people had wanted help answering a question that vexed him: Why were baseball professionals forever attempting to explain essentially random and therefore inexplicable events?

    “Baseball men, living from day to day in the clutch of carefully metered chance occurrences, have developed an entire bestiary of imagined causes to tie together and thus make sense of patterns that are in truth entirely accidental,” James wrote.

    “They have an entire vocabulary of completely imaginary concepts used to tie together chance groupings. It includes ‘momentum,’ ‘confidence,’ ‘seeing the ball well,’ ‘slumps,’ ‘guts,’ ‘clutch ability,’ being ‘hot’ and ‘cold,’ ‘not being aggressive’ and my all time favorite the ‘intangibles.’ By such concepts, the baseball man gains a feeling of control over a universe that swings him up and down and tosses him from side to side like a yoyo in a high wind.”

    It wasn’t just baseball he was writing about, James continued. “I think that the randomness of fate applies to all of us as much as baseball men..."

  12. Brian Burke says:

    Ok. Well, in response to those who believe there is some meaning in 'rhythm' and 'identity', let me pose a hypothetical challenge. We'll choose five tv football pundits and ask them to explain each term. If any one of the five actually says something intelligible, I'll concede the point.

    Coach Herm, what does 'rhythm' mean in football?

    Well, it means. You see. You just gotta. There's this. You play with a rhythm. And that's something an offense needs. See. Without it you just got no rhythm. And that's not good.

  13. James Willoughby says:

    Isn't establishing an identity on offense a sound game-theoretical concept? Isn't it like levelling in poker or hustling in pool, for instance?
    To exploit weaknesses in how your opponent is thinking, it is desirable to push their strategy towards an extreme by giving them clear information as to how you are thinking first.
    Establishing an identity is simply being the dominant party in this exchange.

  14. Jim Glass says:

    Did I see mini-Herm tipping into a cooler in a TV ad? Someone should flick him in to swim with the melting ice and close the top, then store the thing in the basement forever.

    He was the only HC in the last 15 years I would describe as "incompetent". IMHO, FWIW. All verbal posing and blather. More empirically, I looked at historical coaching records compared to what would be expected by regression to the mean and Herm's was the worst of recent decades by a mile. It's a mystery to me how he managed to keep coaching in the NFL for so long, and then pull off a subsequent lucrative career as an "expert" analyst. There ought to be a special award for extraordinary achievement in faking one's way to career success, call it a "Hermie".

    That said, I must admit to believing in "rhythm" in successful play calling even if Herm does too. Give me a 320-pound All Pro OT crushing a 260-pound third-team DE on every play, and I'll have my team running off tackle every 35 seconds like clockwork.

  15. Jeff Fogle says:

    It's not clear whether or not you believe in identity BB. You called it meaningless and nonsensical, then you quickly wrote up a program that did "a better job of identity analysis," which is pretty amazing if it doesn't exist.

    These are common shorthand terms in football and sports in general. If your point was that many announcers don't explain them clearly, there probably wouldn't be much of a debate. If you're saying they don't exist...well, maybe these definitions will help.

    *Identity is not a replacement for the word "good." It's the general term for how teams define what their strengths are and how they try to use them to win games. Some teams are are aggressive, some teams more conservative. Some teams prefer to emphasize the run and short passes as they move the chains to get into position to score. Others prefer to attack deep with a higher risk/higher reward approach. Either style (or others) could be good while representing polar opposites.

    When announcers talk about a team trying to find its identity, it's generally linked to personnel changes on the field because of injuries or new acquisitions. It can also be due to a coaching change where one style is being introduced after another has been dispensed with because it didn't fit the personnel. A team is "between" identities and is trying to find one.

    *Rhythm is not a replacement for the word "good." In football it typically refers to an offensive sequence where a team's success on one play creates an opening for success on the next play...and they continue to do create openings to exploit with successive plays. A short pass to the sideline opens up something in the middle. A pass to the middle opens up a sweep. A sweep opens up a hard-hitter up the middle. Etc...

    An announcer saying "The 49ers need a couple of completions here to get into a rhythm" has it exactly right, because those completions will cause the defense to focus on one thing while the offense then attacks the opening that's created.

    (In hoops, it refers to offensive ball movement that creates open looks, in baseball it's often connected to a starting pitcher moving the ball around in a way that keeps the batter on the defensive).

    Maybe Nigel Tufnel day inspired you to turn the rhetoric up one louder to 11 this week. Announcers, former coaches, former players, and the media in general aren't as stupid as you seem to think. And, talking about voodoo during a week where your ratings finally dropped Philly out of the top 5, finally pushed San Francisco into the top 10, and finally lifted Cincinnati into the top 20 is odd indeed.

    Do you do the voodoo that they do so well?

    These terms are not meaningless just because many analysts aren't adept at explaining them. Have you ever heard a physicist try to explain quantum mechanics?

    You want to blast the media for not being as coherent as they could? I'm guessing most of us are right there with you. Identity and rhythm do exist, and many of your readers know the definitions as you've seen above...

  16. Jim Glass says:

    An announcer saying "The 49ers need a couple of completions here to get into a rhythm" has it exactly right, because those completions will cause the defense to focus on one thing while the offense then attacks the opening that's created.

    That has nothing at all to do with "rhythm". Rhythm definitionally has to do with *a timing pattern* -- "A strong, regular, repeated pattern" the first definition produced by Google, "recurring with measured regularity" the second, "regular recurrence or alternation of different quantities or conditions: the rhythm of the tides", third.

    Can a couple successful pass plays help set up more? Sure.

    But what has that got to do with rhythm? "A series of plays with positive EPA = 'rhythm'" is a rather bizarre use of the word. Just where is "the strong, regular, repeated pattern ... regular recurrence and alteration" in positive EPA plays?

    This is patent nonsense, an Orwellian change of the definiton of a word. Successful plays are successful plays, that's what they are. They set each other up, sure. That's a mark of being *good*, not "rhythmic".

    The only "regular recurrence ... repeated pattern" resulting from the 49ers completing or not completing their passes is in the frequency of their punts.

    But hey, if one really wants to insist on saying a team plays with "rhythm", where is the evidence that this is supposed to be good? Can't a team *suck* with a repeated, steady regularity -- a very consistent rhythm?

    Going three-and-out five times in a row actually does pretty well meet the definition of play having a good rhythm to it.

  17. Connor says:

    When broadcasters say that a certain team has an identity offensively, I think they're trying to say that the team in question understands what its offensive strength is (run or pass) and adjusts its offensive playcalling accordingly. A way you could try and represent whether or not a team has an 'identity' statistically is by multiplying teams' pass/run EPA spread by their Pass % z-scores.

    So, take this as an example: we probably both have an intuition that the Patriots and Broncos both have an strong offensive identity, despite the fact that the Patriots offense is qualitatively different from the Broncos offense (Patriots passing, Broncos Running). To see if our intuitions are correct, we take the Z-score of the Patriots Pass % ([65-59]/5 = 1.2) and multiply that by their EPA spread (.13), giving the Patriots an 'identity score' of (1.2)(1.3) = .156.

    We can do the same with the Broncos ([48-59]/5)(-.09) =.198. Then, if you calculate the z-scores of the Patriots and Broncos identities (.447 and .668, respectively, and yes, I did spend about an hour in Excel this morning), indicating that, while the Patriots and Broncos have markedly different offenses, they both have similarly strong offensive identities (though not extremely strong, which makes sense considering the Patriots and Brady have struggled a bit of late, and the Broncos only transitioned to run-heavy when Tebow became the starter five games ago).

    The good thing about this statistic as it is right now is that running teams that are good at running and passing teams that are good at passing will both have identity scores that are similar in value, meaning that strength of an identity is independent of how it is achieved.

    A bad thing about it, though, is that the average Pass % Z-scores are calculated using the current league passing averages, meaning that it represents a team's identity relative to the current pass/run trends of the league. So, while we may watch a team like the Giants or the Bills and say 'their team has a passing identity', because their Pass % is right around the league Pass %, their identity score will be extremely low. Therefore, if you made a second identity score that used the mean and St. dev. of Pass % for teams over the past 5-10 years, you could have two identity scores, one that compares teams' identities relative to the league as it currently is, and another that compares teams' identities to recent trends in Pass %.

    The advantage of that second stat is that it likely more closely matches our intuitions about a given team's identity, since our intuitions based on our memories of several years worth of games, rather than just the games we have seen this season.

  18. Brian Burke says:

    Jeff-You seem to have a lot more comprehension than I do of the drivel that substitutes for actual analysis.

    I think teams are better at some things than others, and they should usually rely on those better things to move the ball. That is meaningful.

    But the way analysts use the word, there is no underlying meaning other than they're no good at any one particular skill. Simply put, they're "not good."

    But teams can be equally good in running and passing. In fact, in pure theory, all teams should have equal stats in a game theory sense. That doesn't mean they lack an "identity."

    In fact, given opposing defense tendencies and abilities, smart offenses *should* appear to be identity-less because they are adapting to the situation.

    The rhythm thing is bogus. Not buying it. Analysts are just looking for offenses to get a run of successful plays. In other words, "be good."

  19. Jeff Fogle says:

    I think we'll have to disagree about what rhythm means in the context announcers are using it. It's DEFINITELY been used in the context I was explaining...but there are certainly announcers who may have inappropriately broadened that definition of it in a way that mimics their use of "momentum."

    Disagree with Jim Glass that the word only refers to timing and such. It can refer to that. It's often used in the manner I talked about, particularly when you broaden the scope to all sports. (Agree with Jim Glass on a lot of other things though)

    Regarding game theory and appearing to be identity-less (which, again by the way, suggests that Brian DOES believe identity isn't a meaningless word or there would be nothing to avoid), I can see that in a league where everyone is exactly even. But, if a team has strengths that are difficult to stop...then it would seem that they would slant their plays in the direction of what opponents are having trouble stopping. Use enough misdrection to keep them honest...but emphasize the strength.

    Justin Verlander mixes up his pitches to a degree...but he doesn't try to be identity-less as his best stuff is unhittable.

    Oklahoma's defensive athletes had no trouble attacking the Texas backfield because their speed and athleticism trumped any slow-developing trickeration the new-look Longhorn offense could try with its inexperienced quarterbacks. Continuing to attack with that edge would make a lot more sense in that situation than backing off to stay identity-less.

    There's more competitive balance in the NFL...but it's not a league with 32-team parity. There aren't enough star quarterbacks around to lead 32 rosters...which is defacto evidence that teams WITH those star quarterbacks should slant their attack in that direction.

    Coaches don't create identities based on their teams strengths because they're idiots....they do it because it works. Announcers aren't referring to something that doesn't exist or doesn't work...they're referring to the nature of how winning teams are built. There were different identities for the Colts in the Manning years and the Steelers in the Roethlisberger years. Real, tangible differences in the ways they went about doing things to win. Each was good, but understanding their identities brought more context to what "good" meant in terms of what you could expect from the defenses and what you could expect from the offenses.

    That's my take anyway. Hope everyone has a great weekend...

  20. Ian B says:

    First I want to support Brad's mention of Smart Football. The blog does a great job describing how offenses are built and how playcallers use constraint plays to react to what defenses are showing them - the essence of game theory.

    Second I think there's a lot of untapped potential with this type of analysis to look at where teams could adjust for better results. Two other places to look would be sections of the field (short/middle/deep, right/center/left), and also defense. While defense is more reactive than offense, looking at differences in defensive efficiencies could highlight where talent and schemes are being most taken advantage of.

  21. James says:

    I think identity, like a lot of football terms, had an original meaning that had actual value, but it was then copied and overused to the point that it lost all significant meaning.

    So I agree with everyone when they talk about what identity *should* mean, but if you listen to announcers they only ever use that term when things are going bad ("the giants need to get back to their offensive identity"... and the announcer usually has the wrong opinion) or when things start turning around ("the cowboys have finally found an offense identity running the ball"... which is also known as winning).

  22. Odoacer says:

    If you tell a kid he is bad, he will start to act bad as it becomes his identity. People generally become what they believe themselves to be. If you tell a player he is a tough, aggressive type player, he will probably become that if only because he believe it to be true. And importantly, he will probably play that way at a higher level than if he lacks an identity. I see no reason why this individual characteristic should not apply to a group and in fact the additional There are whole schools devoted to psychology. I am not sure why it is so casually dismissed by football quants other than other obvious fact being it is difficult to measure.

    Rhythm is a qualitative way of saying success rate with perhaps an implicit underlying pyschological element to it (confidence, focus?). I believe Brian has proven its value.

    With regards to announcers and writers, that they might butcher or dilute the value of the concept in their attempt to fill airtime is a different issue. One should not confuse the folly of the messanger and his delivery with the validity of the concept.

  23. Anonymous says:

    The rhythm thing is bogus. Not buying it. Analysts are just looking for offenses to get a run of successful plays. In other words, "be good."

    The difference between a run of successful plays and a rhythmic run of successful plays is that the rhythmic plays will appear to have been executed as planned, on time, without much need for creativity or exceptional play. Drives with lucky fumble recoveries, 60+ yard rushing touchdowns, multiple penalties, lots of hurries or extended scrables that result in unexpected big plays-- drives like this may be "good" from a results perspective but not "rhythmic."

    I'm not sure that "rhythm" in this sense has any predictive value, although it's true that other sports often involve deliberate attempts to disrupt an opponent's rhythm and mental state: both basketball and tennis come to mind.


Leave a Reply

Note: Only a member of this blog may post a comment.