Three-Cone Drill: Third-and-Long

Three-Cone Drill: Third-and-Long
Three-Cone Drill: Third-and-Long
Photo: USA Today Sports Images

by Rivers McCown

One of my favorite laments, often uttered by Greg Cosell when he goes on one of the myriad of podcasts or radio spots he does, is "At some point it's third-and-8, and what are you going to do?"

Well, it doesn't seem like a complicated question, right? You try to throw the ball. You try to get it past the sticks.

If we know one thing about NFL coaches though, it's that they aren't exactly acting on rational game theory at all times. If the fourth-down revolution actually is over, as Mike Tanier declared, maybe this can be the new front. As a fan of a team that ran nearly nine percent of all third-and-long run plays last season, I can tell you that conservatism is not an uncommon theme in these situations. Teams often do decide that playing it safe is the best course of action.

I decided to dig into our 2012 charting database (we haven't put together 2013 yet) to have a look at how often this happens, and try to project exactly what some certain coaches are protecting against.

Unfortunately, the scope of the project is such that it will take me a few weeks to dig through the numbers to my satisfaction. So, this week we'll just start by defining how often this happened in 2012.

I decided to split these into two separate columns: third-and-7 (third-and-7 or longer) and third-and-10 (third-and-10 or longer). My theory is that coaches would be more willing to run up the white flag if they a) hadn't seen much forward progress on this attempt at a first down or b) were put in a spot where their quarterback wouldn't be able to throw a fast-developing route like a quick slant or curl to pick up a first down.

Coaches on Third-and-Long, 2012
  Third-and-7 Third-and-10 Third-and-7 Success Third-and-10 Success
Plays 3289 1934 26.4% 21.3%
Rushes 257 174 10.5% 6.3%
Screens or Dumpoffs* 554 359 15.1% 10.8%
Other Passes 2074 1178 34.3% 29.2%
6-32 Other Passes** 1802 1043 33.6% 28.4%
*- All passes that travelled less than two yards in the air.
**-Other passes by team without a top-five passing offense in 2012

(Note that the play count does not add up exactly because of quarterback scrambles, aborted snaps, and other plays that aren't as common. We'll explore scrambles next week.)

That's a very aesthetically pleasing curve of slight decline.

I chose two yards as the pass distance because it covered most screens. This count does not include every dumpoff pass or swing pass to a running back -- it just covers those within two yards of the line of scrimmage. I figure at that point you're basically relying on your back or receiver to shed someone to get a first down.

I added a row without the top-5 offenses because I figured it's easy to say a coach should be more aggressive, but harder in practice when you don't have a Peyton Manning or Tom Brady (2012 vintage? Are we at the point where we need to specify that? Yikes). Our top-5 pass offenses last year were New England, Green Bay, Denver, San Francisco, and Seattle. Removing them changed ... almost nothing. A very slight decline.

So in the macro, NFL coaches have spent nearly a quarter (24.6 percent) of their third-down plays in 2012 on what would appear to be sub-optimal decision making. Every screen or draw they run lowers their success rate by a little over 20 percent.

Next week: What are the actual strategic reasons to play it safe? We'll look at trends in sacks, hurries, and interception rate on third-and-long.

Money for Nothing

One common theme that I seem to see pop up every couple of months or so is the idea, from veteran journalists or writers, that writing should not be free. I was re-directed to this idea last week via the Twitter, when Bleacher Report's Ian Kenyon led me to this article posted in the New York Times.

I think I take a somewhat more nuanced stance on this matter than some. Look, the obvious conclusion is that if you have any interest in becoming a writer in an industry as popular as sports, you're going to have to do some writing for free. Much like, say, asking someone to mentor you, you are at some point going to have demonstrate actual interest and intent in your field before anybody believes that you are serious. And yes, that has to be self-motivated interest. That is going to be actual work, whether you are passionate about a field or not. Amazing how quickly people fade away from their dreams once those chips are down.

I don't think the question for a beginning sports writer is whether to work for free or not -- the question is where you begin to write for free. And it's a question that will have different answers for different people.

The calculus is not insanely complicated, I promise. Mass media sites that don't pay for the majority of their content are built on the simple realities of the market. The good news for them is that because so many people want to write about sports, it's not hard to find someone to do it for free. The bad news is that the vast majority of people who start out at that level are never actually going to be writers -- they'll either burn out and quit, write so poorly they'll be asked to stop writing, or tread along as a replacement-level writer until they see things aren't going anywhere -- by definition, most of that stuff just isn't ever going to be material that sells beyond the SEO perspective.

So the question then becomes one of exposure and one of credibility. If you are, say, an ex-NFL player. Someone who used to work in a front office. Someone who can mix knowledge of coverages with film breakdowns. Someone who can outperform 90 percent of other writers and knows how to fill a niche that hasn't really been touched on yet. If you are one of these people, you have a product that sells itself. All you need to do is get in touch with the right people to make sure your stuff is seen. (And, here's the sneaky part: you probably maintain more of your own identity and personality with your own site than you do by joining a network. As long as things are free, anyway.)

[ad placeholder 3]

The people who benefit most from working on a big media site for free, in my view, are the people who need the most guidance and shaping as writers. The exposure to internet commenters is going to teach them a lot about themselves. Being able to ask for advice from their immediate superiors is going to teach them a lot about the actual art of writing and "web journalism." Being under the same tent with more well-known writers is going to steer them closer to what they do best through imitation and osmosis.

I have a little bit of experience in both worlds. I started writing on my own blog as a game charter in 2009, and I got asked to join (and accepted) and invitation to an SB Nation blog after that season. I was always a very hard worker, and while I definitely improved on my writing over the years, I always have had the ability to put an original twist on a phrase. But joining SB Nation's Texans blog opened me up to a lot of new opportunities: much more attention than I'd ever received was chief among them. I was a lousy self-promoter. And, if I'm being honest with myself, I'm still not very good at it.

I learned I was probably not going to be a good beat reporter without a lot more work when I was allowed to attend the NFL Combine under the SB Nation window in 2010. Everyone was asking questions that were either monotonous, borderline stupid, or completely uneventful. I decided to challenge that status quo. So -- I thought about this a lot on the airplane to Indianapolis -- I decided the best way to do this would be to start a running collection of the weirdest question a player had been asked by a team or media member.

If you've never been to the combine before, each prospect of non-renown (i.e. not a projected first-rounder) gets a little table to talk to media for 10-15 minutes. I got a pretty good quote from future Chargers nose tackle Cam Thomas the first time I sat down and asked a question, so I was feeling good about myself. Next was future Bills end Alex Carrington. I asked him the question, and he just stared at me stone-faced for about five seconds.

"Probably this one."

My face became a tomato and I spent the rest of the day picking up the shattered pieces of my psyche. I failed. Spectacularly. But I got the opportunity to fail spectacularly. And that was what mattered as far as figuring out what niches of this whole sports writing thing I'd actually be good at.

Do I as a writer feel like we should all get paid? Sure. I work my ass off. And I think it's fair to lament that the profession is not one that is economically valued by our society anymore. But the economic reality of the situation is that unless you've got a perspective or talent that few have, you're going to need to work your way up the chute. I know it's sort of blasphemy to compare the two, but these modern-day mass media sites where you can write for free ... they're sort of like writing colleges you don't have to pay for. Most people aren't going to come out of them with a flashy diploma and a job in sports, just like 90 percent of the people I took creative writing classes with aren't actually writers post-college. But if you've got your head on straight, and you're willing to see how far it can take you, you'll learn a lot about yourself with some of the opportunities you get.

Lofty Nuggets

by Andrew Potter

Atlanta at Carolina -- 1:00 PM ET (FOX)

A divisional matchup between two franchises which are trending in opposite directions gets us started this Sunday afternoon. The Carolina Panthers are in the ascendancy: their 4-3 record speaks more to their poor late-and-close game management than their general play, and their 96 points allowed is the fewest in the NFL. The Atlanta Falcons, by contrast, are collapsing under the weight of injuries and poor offensive line play as Matt Ryan valiantly battles to construct a productive pass offense from the duct tape and bailing wire that is Drew Davis (53.0% DVOA, nine targets), Darius Johnson (-22.5%, 12 targets), and Levine Toilolo (-12.2%, nine targets). With Julio Jones out for the season and Roddy White questionable for Sunday's game, look for the Panthers to built a small fort around Tony Gonzalez (-7.1% DVOA, 59 targets) and dare the Falcons backups to beat them. Carolina's offense (12.8%) should also have its way with an Atlanta defense (15.6%) which is likewise starting backups at key positions, leaving this pair of franchises to continue along their respective, divergent paths.

Minnesota at Dallas -- 1:00 PM ET (FOX)

[Vikings Quarterback Flavor of the Week] leads an offense which doesn't travel well (-20.5% DVOA in road games) into Dallas to face a defense which, while average against both the pass (8.2%, 16th) and the run (-8.0%, 17th), has struggled at home this season -- its 4.7% home DVOA ranks 26th. The potentially explosive Dallas offense has cooled in the past few weeks -- negative DVOA every week since the Denver shootout -- and the loss of Brian Waters for the season adds another potential issue to what is usually the better of the Cowboys' units. Minnesota's defense is poor, however, and Cowboys quarterback Tony Romo (18.6% DVOA, 615 DYAR) is not some generic passer devoid of weapons: Dez Bryant (12.3% DVOA), Terrance Williams (15.3%), and Jason Witten (6.0%) are all reliable targets even in the absence of injury-hit Romo favorite Miles Austin, and all should have plenty of opportunities to succeed against Minnesota's 26th-ranked pass defense (-18.5%).

New Orleans at NY Jets -- 1:00PM ET (FOX)

Cincinnati's dismantling of the Jets pass defense last week was the clearest sign yet of the decline of that unit this season. Under Rex Ryan, the Jets pass defense has had a negative DVOA (a good thing for a defense) every single year, never finishing outside the top ten by DVOA (they finished exactly tenth last year). This season, their 11.6% pass defense DVOA ranks 21st, and they've only had two of eight games with negative pass defense DVOA (Buffalo in Week 3 and New England in Week 7). This week's opponent, New Orleans, hasn't had a negative passing performance since Week 2 against Tampa Bay, and has been above 25% DVOA for every game except that one. New York has the league's best run defense (-31.1%), but against this Saints team that's unlikely to matter if they can't stop the pass. If forced into a shootout Geno Smith (-25.6% DVOA, 34th) and this Jets pass offense (-18.4%, 31st) aren't likely to come close to keeping up against Rob Ryan's improved Saints pass defense (-6.6%, seventh).

Tennessee at St. Louis -- 1:00 PM ET (CBS)

Jeff Fisher's old team visits Jeff Fisher's new team in a game featuring two franchises who were considered dark horses for the playoffs earlier this year. Those hopes were based on strong defense and possible offensive improvement, as their young quarterbacks were expected to benefit from another year of professional experience. While both quarterbacks did show improvement, both have also missed time due to injury -- Sam Bradford won't take another snap in 2013 -- and neither defense has quite lived up to its promise leaving both with three-win losing records. For Tennessee, that record actually has them in a decent position from which to make a postseason run -- playoff odds of 17.8%, whereas the odds for the Rams are 0.1% -- and the Titans are the better side by DVOA on both offense (-3.1% versus -13.6%) and defense (-1.3% versus 3.7%). If they're to realize those playoff hopes, games like this against inferior opponents must show up in the win column on Monday mornings.

Kansas City at Buffalo -- 1:00 PM ET (CBS)

Kansas City's 192 points scored is the lowest total for an 8-0 team in the history of the NFL: the most common record for teams with 192 points at this stage of the season is 6-2, and the 2007 Patriots had a higher points differential than the 2013 Chiefs' total points scored. 6-2 is also the most common record for teams which allow 98 points in their first eight games, so the Chiefs are an identically placed outlier in terms of both points for and points against. They travel this week to face a Buffalo team which may be starting its third quarterback of the season -- and hey, that's nothing new for Kansas City. They barely beat a Cleveland side which was also onto its third quarterback. Kansas City's pass defense is legitimately good, but their win/loss record probably has more to do with their schedule than their play. Buffalo's pass defense is also legitimately good, so this is likely to be the type of field position trench fight in which Kansas City's huge advantage returning punts (7.2% on punt returns versus Buffalo's -10.7% on punts) could be one of the deciding factors.

[ad placeholder 4]

San Diego at Washington -- 1:00 PM ET (CBS)

Matt Waldman analyzed what's wrong with Washington's offense this week, a frequent theme among scouts and bloggers who've covered their games this season. In theory, this week provides a good chance to get that offense back on track; despite Jacksonville being ranked 32nd in both pass and run defense, San Diego has the worst overall defensive DVOA at 22.8% (this happens quite frequently and is caused by run/pass play distribution). Instead, Washington's 27th-ranked 8.0% defense is the much bigger concern for this game, as it matches up against San Diego's third-ranked 25.0% offense. Philip Rivers is second only to Peyton Manning for efficiency (42.3% DVOA) and is closer to Manning than third-ranked Drew Brees by that measure. Robert Griffin (-15.2%), by contrast, is sandwiched between Christian Ponder and Eli Manning: a ranking that tells its own story about the productivity of Washington's hobbled young star.

Philadelphia at Oakland -- 1:00 PM ET (FOX)

Nick Foles returns for this game, sparing Eagles fans the Matt Barkley experience at least for the opening few plays. Two weeks ago, Foles was the most efficient quarterback in the league by DVOA. Now, he's seventh and falling, and the collapsing Eagles offense -- averaging -50% DVOA the past two weeks amid a rash of quarterback injuries, though still ranked first in rushing -- is about to visit an Oakland side whose defensive DVOA is much better at home (-6.5%, 12th) than on the road (1.8%, 29th). Oakland's offense, conversely, is much less bad on the road (-9.9%) than at home (-22.5%), Terrelle Pryor record-setting runs notwithstanding. Oakland's offensive home/road split is good news for a struggling and inconsistent Eagles defense which has been particularly awful on its travels (19.5%, 31st), while two of the league's worst special teams units (Oakland's ranks 26th; Philadelphia's ranks 28th) complete the profile of a game which is more likely to affect draft position than postseason participation.

Tampa Bay at Seattle -- 4:05 PM ET (FOX)

The last time Tampa Bay had a top-three draft pick was 2010, when they selected Defensive Tackle Gerald McCoy. The Buccaneers won ten games that year, have won 11 games in total since, and are odds-on to make another top-three selection this year (73.9% chance of a top-three pick, 16.8% chance of picking first overall). They take their -9.3% pass offense on the road to face the league's best pass defense (-32.5%) with their opening-day starting quarterback cut, their star running back badly injured, and one of their starting wide receivers on injured reserve. It's been that kind of year for Buccaneers fans. Seattle's offense has injury problems of its own, and Tampa Bay's third-ranked run defense (-18.7%) is a terrible matchup for an offense with problems on the line and a banged-up running back (Marshawn Lynch, -2.0% DVOA), but it's hard to see how Seattle's defense doesn't outscore Tampa Bay's offense by itself, leaving Lynch and company to simply avoid mistakes.

Baltimore at Cleveland -- 4:25 PM ET (CBS)

Cleveland, already on its third starting quarterback of the year, hosts a Baltimore side which still manages to be worse than the Browns on offense by overall DVOA (-14.5% versus -11.4%) despite being settled at the game's most important position. That's primarily because Baltimore's running game (-30.2%) has been terrible this year -- neck and neck with the league-worst New York Giants (-30.7%). The Browns defense is, like most defenses in 2013, competent against the run (-4.0%, 21st) but vulnerable to the pass (8.9%). It is, however, very good against No. 1 receivers (-20.8%, sixth) so the Ravens will need the likes of Marlon Brown (6.7% DVOA on 34 targets) to be productive across from clear number one Torrey Smith (20.0% DVOA on 61 targets). Cleveland's offensive prospects are as difficult to predict as its quarterback situation, but Jason Campbell's 122 DYAR ranked sixth in Week 8. Against a good Ravens defense, a similar ranking next week would be a very good sign for the Browns signal caller, albeit another clear indication for the franchise that opening day starter Brandon Weeden is part of the problem, not part of the solution.

Pittsburgh at New England -- 4:25 PM ET (CBS)

It's been five years since Pittsburgh last played in Foxboro in a regular season game. Matt Cassel was the Patriots quarterback in November 2008, and the Steelers won 33-10. Against Tom Brady, it's been a different story historically for the Steelers, but this is not your vintage Tom Brady: his -16.6% DVOA ranks 29th out of 39 qualifying passers. Ben Roethlisberger is also quietly having a down year: he hasn't finished with a negative DVOA since 2008 -- coincidentally, the last year the Steelers played in Foxboro -- and is currently ranked 18th by both DVOA and DYAR. All of that means those who watch this game hoping for an epic battle of multiple Super Bowl champion quarterbacks may be sorely disappointed: the quarterback matchup numbers would look almost identical for Alex Smith against Chad Henne. The Patriots are better than the Steelers on defense (-6.2% versus 0.8%) and have the league's best special teams (7.4%), so a Patriots win remains the most likely outcome -- but if Brady's unbeaten home record against the Steelers continues, this time it may well be in spite of him rather than inspired by him.

Indianapolis at Houston -- 8:30 PM ET (NBC)

A second divisional matchup between two franchises which are trending in opposite directions bookends the Week 9 Sunday slate. Case Keenum continues at quarterback for the Texans after his relatively composed debut (15-of-25, 271 yards, 1 touchdown, no interceptions, 38 DYAR), but his presence in the starting lineup is one indicator of how quickly the team around him has gone from possible Super Bowl contender to likely playoff outsider -- our playoff odds projection currently gives the 2-5 Texans just a 2.5% chance of reaching the postseason. Indianapolis, by contrast, reaches the postseason in 96.6% of simulations, and wins the division 95.4% of the time -- the highest figure in the league for a division win. That makes this home game critical for Houston's season, but it comes the only team in the league which is top ten by DVOA on offense (17.1%, fifth), defense (-4.4%, ninth), and special teams (4.6%, ninth). If the turnaround in Houston's fortunes since 2011 is drastic, the turnaround in Indianapolis is much more remarkable -- and has much broader implications for the rest of the AFC.


6 comments, Last at 04 Nov 2013, 8:38pm

1 Re: Three-Cone Drill: Third-and-Long

"So in the macro, NFL coaches have spent nearly a quarter (24.6 percent) of their third-down plays in 2012 on what would appear to be sub-optimal decision making. Every screen or draw they run lowers their success rate by a little over 20 percent."

Here's the problem - DVOA will never be able to find the underlying decision-making behind low-efficiency 3rd down calls, because DVOA is a per-play metric. DVOA is simply looking at the wrong decision making concepts to find this. You really need something more like a per-drive metric, or a field-position value metric. Those are able to comprehend why the decision would be made and be able to reflect whether the choice was valued added or subtracted.

It's not so much a flaw in DVOA as it is a decision made on an orthogonal axis that DVOA doesn't measure.

2 Re: Three-Cone Drill: Third-and-Long

Do you think the higher success rate from throwing is due to the fact that it is more difficult to stop a pass play, or that by running or throwing screens keeps the defense from knowing with certainty that you are going to throw? I assume it is probably a combination of the two. Put differently, do you think if teams stopped calling sub-optimal plays the success rate for passes would drop significantly?

5 Re: Three-Cone Drill: Third-and-Long

The thing is that it doesn't really matter. From a game theoretical stand point, to the best of my understanding, the offense is in control here and should keep calling pass plays, forcing the defense to back off the run and play the pass harder until the two are of essentially equal value.

I think one major question is how often the run was intrinsically valuable in these situations, either because you were in a four down offense and the possibility of creating fourth and short was considered more valuable than the extra 20% success rate, or because the team is sitting on a small late game lead and the extra 45 seconds. Those may account for a lot of the runs.

The only explanation that comes to me for all the dumpoffs are all downfield receivers being covered and Manning/Brady types pulling a "You're more likely to get 7 yards than I am."

6 Re: Three-Cone Drill: Third-and-Long

From a game theory standpoint you'd pass downfield more, forcing defensive adjustments to this, until the win-expectation shift became equal, not the conversion rates.

But win-expectation shifts vary based on situation. Passing downfield is riskier in specific ways. First, there's a greater risk of a turnover; second, there's greater risk of a sack. In some cases this matters a lot more than in others.

When we decide whether to go for a two-point conversion, we ask whether we care more about the difference between 0 and 1 point, or the difference between 1 and 2 points. If we're up 1, being up 2 isn't much of an improvement but being up 3 is, so we're more likely to go for 2. If we're tied, getting up by 1 is really important and getting up by 2 doesn't matter so much, so we're very unlikely to go for 2.

The same question applies for third-down conversions. Are you in easy field-goal range? Drop back and take a sack before a pick. Hard field-goal range? Avoid the sack, have at least a quick outlet. Are you planning to go for it on 4th if you come up a couple yards short? Then you'll be more willing to settle for a play that "sets up a manageable 4th down". Sometimes the difference between 0 and 3 points matters more than the difference between 3 and 7, sometimes less.

It wouldn't surprise me if defenses couldn't sacrifice ability to stop the draw on 3rd-and-long in order to stop downfield passes much more than they already do -- that is, that 11 professional defenders, even if they're all assigned to pass coverage or pass rush, are good enough to stop a draw or screen before it gains 7 yards most of the time.

3 Re: Three-Cone Drill: Third-and-Long

Isn't 3rd and long going to be massively affected by selection bias? The teams who believe they're going to convert are the ones trying long passes. The teams who don't are the ones using give up draws.