2018 Offensive Personnel Analysis
by Bryan Knowles
Here at Football Outsiders, we've been tracking offensive personnel usage since 2010. The dominant story of this decade has been the rise of the three-wideout set. Just as the 1940s are known for the T-formation killing the single wing and the 1980s are known for the rise of the West Coast offense, the 2010s will go down in the Great Book of NFL Offensive Thingamajigs as the decade when the three-wideout set became the default offensive mindset. When we started tracking this in 2010, teams used 11 personnel (one running back, one tight end, three wide receivers) less than 40 percent of the time; it was the most prevalent personnel grouping, but that was mostly because a large chunk of teams were still split between favoring two backs or two tight ends, with three-wide sets being for passing downs alone. Nowadays, taking that third receiver off the field is the odd move. 11 personnel is not only the default offensive formation, but is also the most effective one by a significant margin. In general, offensive diversity isn't based on the personnel you put on the field anymore; it's just about where you line them up and how you use them.
Before we get into things any further, a quick note: this is personnel data, not formation data. When Tarik Cohen goes out wide, he's still counted as a running back. When Cordarrelle Patterson lines up in the backfield, he still counts as a wide receiver. We're using the standard numerical system where the first digit is the number of backs and the second digit is the number of tight ends. We count formations with six or more offensive linemen separately, rather than counting the offensive linemen as tight ends. A formation with two backs, one tight end, one wideout, and six offensive linemen is marked as "621" and not "22." Last year, we noted that 11 personnel usage dropped for the first time since we started recording; a league-wide shift, albeit of just over one percent. Had we reached equilibrium? Had we finally reached the end of the era of 11 expansion? It was something we promised to keep an eye on going forward. Well, we kept an eye on it, and the answer was "no."
|Offensive Personnel Groupings|
|Personnel||2017 Pct||2018 Pct||Difference||2018 DVOA|
After the one-year blip, 11 personnel usage is back on the rise, setting a new record by being used on 64.2 percent of all offensive plays in 2018. Twenty-one teams saw their 11 personnel usage rise from 2017; 18 of them saw it rise by at least five percent and 11 saw double-digit increases. Thirty-one teams used 11 personnel more than half the time, with the Rams setting a new record by using it on 92.3 percent of their snaps. Every other formation dropped off, barring a couple used on one or two plays. If this keeps up, we're going to have to change the focus of this yearly article, because differentiating between personnel groupings is rapidly becoming a thing of the past. Other teams in the past have used three-wide sets on nearly 90 percent of their plays, most notably Ben McAdoo's Giants, but we've never had one team so slavishly devoted to one personnel grouping as Sean McVay's Rams. In fact, Los Angeles only had five different personnel groupings all year long:
- 11 personnel, used on 958 plays.
- 12 personnel, used on 70 plays. Fifty-six of those plays came in Weeks 16 and 17 when Todd Gurley was out with a knee injury.
- 10 personnel, used on four plays.
- 01 personnel and 13 personnel, used on three plays each.
Of course, this lack of creativity and overly predictable usage of his players was a disaster for the Rams offense, as they ended up finishing ... uh, second in offensive DVOA, rising from 11.1% in 2017 to 24.6% last season. With success like that, who needs a second tight end? Until Gurley went down, the Rams were using 11 personnel 97.5 percent of the time; that may be the number to shoot for in 2019. The Rams were one of the 11 teams to see their 11 personnel usage go up at least 11 percentage points between 2017 and 2018, but just barely. The biggest increase -- 20.8 percent -- went to Carolina, with Norv Turner replacing Mike Shula calling plays. In fact, six of those 11 teams had new playcallers -- Byron Leftwich and Mike McCoy in Arizona, Matt Nagy in Chicago, Brian Daboll in Buffalo, Matt LaFleur in Tennessee, and John DeFilippo in Minnesota joined Turner. When teams are looking for an offensive spark, they're looking more and more towards three-wide playcallers, because that's now the standard. As we see more and more first-time playcallers get jobs, we could see 11 personnel usage climbing higher and higher as more of the old guard is rotated out (or, in the case of Turner, old dogs learn new tricks).
|DVOA in 11 Personnel|
|Team||11 Pct||Rk||Yds||11 DVOA||Rk||Non-11 DVOA||Rk||Diff||Rk|
Thirty-one teams are playing football in the 2010s, and then we have the San Francisco 49ers, still out there trotting out fullbacks like it was 1974. They stick out like a sore thumb there at the bottom of the table. We had eight teams under 50 percent two years ago, but the other seven teams all got with the program in 2018. The 49ers were the only team in the league to not use 11 as their primary formation; they didn't have any formation hit 50 percent, but the plurality of their plays came in 21. This is what happens when you have the highest-paid fullback in the league on your roster, I suppose. But you know what? They were right to do that, considering their available talent.
In 21 personnel, their offensive DVOA rose to -1.5%, and their 12 personnel groupings were at -6.2%, both higher than what they got out of 11 personnel. This is a case of "get better receivers" more than anything else, but at least the 49ers correctly responded to their lack of receiver depth in 2018. When Kendrick Bourne is your top-targeted wideout, you probably don't have three healthy NFL-quality wideouts to put on the field at any one time.
The same can not be said for Arizona, Buffalo, and Washington, the bottom three teams in 11-personnel DVOA. While there were likely no strategic decisions that could have made any of those offenses good, considering their quarterback situations, they at least could have mitigated some of the damage. All three were significantly better when going away from 11 personnel, and yet all three were among the ten teams that used it the most. Arizona, in particular, was poorly conceived. In a world where having more than two receivers on the field was illegal, Arizona would rise all the way from last in offensive DVOA to … OK, just 29th, but going from -41.1% to -16.0% is a massive improvement. I suspect a significant amount of that improvement came from having Jermaine Gresham in to pass-block, as the Cardinals' offensive line was a disaster last season. Continuing to do something that was clearly not working when other options were available is one of the main reasons Arizona's coaches from 2018 are all in different places in 2019.
On the flip side, there were a few teams that could have used 11 personnel a bit more often. The Patriots and Ravens started going that way after years of favoring two tight-end sets; they were two of the teams who increased their 11 personnel by double digits. For the Pats, that was probably sparked by the aging of Rob Gronkowski, who has meant that New England really hasn't needed an extra wide receiver threat for years. It's not just Gronk, though; Dwayne Allen played fewer snaps than New England second tight ends typically do; the Patriots ran just 116 plays with two tight ends, down from 253 in 2017. With Gronk gone, expect that trend to continue, especially if Josh Gordon is reinstated.
As for the Ravens, you'd think that a team that used the least 11 personnel in 2017 and drafted two tight ends early in the 2018 draft would continue to be near the bottom, but no; they shot up dramatically. This is what happens when you bring in Willie Snead, Michael Crabtree and John Brown to replace Mike Wallace, Jeremy Maclin and Breshad Perriman. It's worth noting that there was not a significant difference between Joe Flacco's Ravens and Lamar Jackson's Ravens in personnel usage. It'll be interesting to see if their 11 makeover continues now that Greg Roman is taking over playcalling duties, Roman's offenses have historically focused around tight ends and fullbacks rather than splitting receivers out wide.
Then you have the Saints, who, for the second year in a row, ranked near the top in 11-personnel DVOA, and near the bottom in 11-personnel usage. They have consistently been a better offense with three wideouts on the field, but they really throttled back on those sets starting in 2017 -- or, to put it another way, just as Alvin Kamara showed up. In 2017, the Saints DVOA split between 11 and 12 personnel was nominal, less than two percent, with 21 personnel less than ten percent behind that. While subbing in that extra tight end or running back made the Saints less efficient, it wasn't by much. Last season, however, the DVOA gaps widened, and now we have something to talk about.
|Saints Personnel Usage|
|Personnel||Frequency||DVOA||Pass%||Pass DVOA||Run DVOA|
Most of the difference between the Saints' three-wide and two-tight end sets are just how often they choose to run out of each. They're about equally good at passing out of both formations, but much better at running without the extra tight end on the field. This is not a Saints-only problem; the league had a DVOA of 3.9% while running out of 11 and a -13.3% DVOA running out of anything else. Bringing extra bodies onto the field not only telegraphs a run, but clogs up running lanes. But the Saints' struggles with both Kamara and Mark Ingram on the field are real; they were lagging behind in 2017 and were much worse last season. With Ingram out and Latavius Murray in, perhaps we'll see fewer situations where both Saints running backs hit the field simultaneously. Other tidbits from our personnel files:
- While most of the league saw their 11 personnel go up, Philadelphia and Detroit went screaming in the other direction, losing more than 10 percentage points from 2017 to 2018. For the Lions, trading away Golden Tate meant they ended up using more of Levine Toilolo and Luke Willson; the Eagles' injury issues at receiver meant they had to trade for Tate midseason. The teams used 11 personnel 60.6 percent of the time with Tate on their roster, and 54.0 percent of the time without him.
- The 49ers used two running backs on 489 plays, with New England behind them at 378. Outside those two, only Denver and New Orleans had even 200 two-back snaps. The Rams didn't have any! RIP pro set; we hardly knew ye.
- The Rams were also the only team to have no plays with more than five offensive linemen on the field at any one time. The Packers nearly joined them, but ruined their chance at pointless statistical tomfoolery by running one play against New England with an extra lineman, backed up on their own 1-yard line.
- As further evidence that Pete Caroll and Brian Schottenheimer live on a different world from the rest of us, the Seahawks alone were responsible for 21.3 percent of all the six-plus-linemen sets last season. They ran 66 plays in 613 (six linemen, one running back, three tight ends), which is about the biggest set you can possibly get onto the field. That alone would have been the third-most six-linemen sets in the league, even without all of their 612 (six linemen, one running back, two tight ends). Don't knock it if it works, though; Seattle's DVOA out of the 613 was 15.3%.
- Perhaps in an attempt to show the world just how little they needed Le'Veon Bell, the Steelers ran 99 plays with no running backs on the field, 31.9% of the league's total.
- Excluding the Rams, the first "missing" formation in the league was 13 personnel. While it was the fourth-most common personnel grouping in the league last year, neither the Chiefs nor the Seahawks ran a play out of it. A third tight end is just a missed opportunity to use an extra lineman for Seattle, and Kansas City was too busy trying to set offensive records to try to figure out how to get Deon Yelder onto the field.
5 comments, Last at 01 Jul 2019, 7:31pm
#1 by Harris // Jun 27, 2019 - 4:43pm
There is a very good chance the Eagles use a lot more 12 this season as they phase out 2020 free agent Nelson Agholor and look to capitalize on the talents of Dallas Goedert, who proved to be a good receiver and surprisingly effective blocker as a rookie.
#2 by herewegobrowni… // Jun 27, 2019 - 10:01pm
"But you know what? They were right to do that, considering their available talent. In 21 personnel, their offensive DVOA rose to -1.5%, and their 12 personnel groupings were at -6.2%, both higher than what they got out of 11 personnel."
Was there significant variance in which groupings were used more with JimmyG and which were with lesser QBs, though?
#3 by Bryan Knowles // Jun 28, 2019 - 3:01pm
With Jimmy Garoppolo:
11: 37.6 percent of plays; -34.5% DVOA
21: 43.5 percent of plays; 12.4% DVOA
12: 9.1 percent of plays; 35.4% DVOA
With C.J. Beathard:
11: 37.5 percent of plays; -15.7% DVOA
21: 40.2 percent of plays; 4.6% DVOA
12: 11.0 percent of plays; -58.7% DVOA
With Nick Mullens:
11: 40.2 percent of plays; 0.1% DVOA
21: 42.7 percent of plays; -11.3% DVOA
12: 10.2 percent of plays; 17.4% DVOA
It's interesting that, while the 49ers were better offensively during Mullen's tenure, they were better at their primary formation with Garoppolo under center. A lot of that, I think, comes from the receivers available early in the year to Jimmy G. He only started three games; Marquise Goodwin missed one and a half of them and Dante Pettis had not climbed his way up the depth chart yet.
Also, Mullens is way better than Beathard, just, in general.
#4 by Jose21crisis // Jul 01, 2019 - 5:32pm
If memory serves, Pittsburgh ran a true 5WR set a few times (I remember that 5 wide set against NE, part of those 99 plays without RB). The 00 set is pretty rare, but how rare was the 00 set this season?
#5 by Bryan Knowles // Jul 01, 2019 - 7:31pm
Rare, but not unheard of -- just 38 snaps of 00 last year, and 29 of those came from Pittsburgh. Dallas and Buffalo had four a piece, and Atlanta had the other one.
Pittsburgh also had a league-high 68 snaps (out of 212 leaguewide) in 01, which is the same thing but with Vance McDonald out there instead of Eli Rogers.