2021 Defensive Personnel: Cowboys and Bills Abandon the Base
NFL Offseason - After looking at the offensive side of the ball earlier this week, it's time to flip things around and look at the most commonly used personnel groupings on defense. And, like we say every year, that means we get to talk about how the league has nickel and dimed "base" defense out of existence. We saw a new high-water mark for nickel defense in 2021, with defenses coming out with five defensive backs on 61.4% of all snaps. Base defenses are now used on less than a quarter of all plays, making the name more misleading than ever before.
It's a natural response to the increasing number of three-wide sets we see around the league, and it's just the latest step in a process that has been going on for over a century. The big question of the 2020s is how to deal with so much speed on the field, making your defense more mobile to deal with the increased power of the passing game. That was as true in the 1920s as it is in the 2020s, as the history of defensive football is a fairly straight line pulling more and more bodies out of the box and into more versatile locations deeper on the field. The question we have now is where does this stop? Nickel defensive usage had plateaued a bit over the past couple seasons, hovering just under 60%, but now it seems to have started forward momentum again. How much nickel can a team realistically run? And who are the few teams pushing back against the nickelization of the NFL? Let's dive in.
|Defensive Personnel Groupings, 2021|
|Personnel||2020 Pct||2021 Pct||Difference||2021 DVOA|
A few quick notes before we continue:
- For the record, 57% of base snaps were in 3-4 and 40% were in 4-3. Most of the rest were 2-5s, used occasionally by the Chiefs, Cardinals, and Falcons. The lead between 3-4 and 4-3 tends to bounce back and forth from year to year as coordinators shift.
- There was less front-mixing in 2021 than we had seen in previous seasons. Twenty-two teams were in either 4-3 or 3-4 on at least 75% of their base snaps. That's up from 14 in 2020, and I have no explanation for why there was such a dramatic shift. It reverses a trend we have seen over the past few seasons, dropping from 25 to 22 to 14 since 2018. Perhaps 2020 was just a strange, COVID-fueled outlier.
- "Dime+" includes any package with more than five defensive backs. That includes all your dime packages, as well as the 128 snaps in quarter and 12 snaps with eight defensive backs on the field. We even had one play with nine defensive backs (gold dollar defense?) with the Titans playing an uber-prevent against the Seahawks near the end of regulation in a tie game in Week 2. Russell Wilson was able to scramble for 12 yards and a first down, but it ate up nearly the rest of the clock and still kept Seattle out of range for even a Hail Mary attempt, so all's well that ends well.
- All defenses with fewer than four defensive backs are either "Big" or "Goal Line." "Big" includes anything with three defensive backs and fewer than four defensive linemen (mostly 4-4-3 and 3-5-3), while "Goal Line" is everything else. Despite its name, there's no requirement for goal-line defenses to actually be played on the goal line; Arizona had five snaps of 4-5 on the other side of the 50 in Weeks 1 and 2. Lining up in 4-5 against trips on first-and-10 is a bold decision, forced by injuries, but both the Vikings and Titans ran into the 4-5 on all five plays.
A 1.7% bump in nickel rate is not exactly a paradigm-shifting change. We're talking one or two extra snaps of nickel per team per game; not something to be writing home about. But this is the second year in a row where nickel has gone up while both base and dime have gone down. Nickel usage went up in every split—against 11, 12 and 21 personnel; against the pass and against the run; you name it. Nickel is the Swiss army knife of defenses, and shows no sign of stopping anytime soon. The league didn't set the record for most defensive backs per play, dropping from 4.89 to 4.88 as dime+ packages faded slightly, but that's a rounding error of a difference. Nickel is everywhere.
Defenses Against Top Personnel Groups
|Defensive Personnel DVOA Breakdown, 2021|
|Split||% Used||DVOA||% Used||DVOA||% Used||DVOA|
While base defense continues to have a better defensive DVOA than nickel, that's mostly because of play selection. Base is being used more and more in obvious running situations, and since rushing is less efficient then passing, base defense looks better and better as more of its sample is taken up by run plays. So yes, if you're sure the opposing team is going to run, it's better to have an extra lineman or linebacker in the box. In obvious rushing situations (say, third or fourth down with 3 or fewer yards to go), base outperforms nickel against the run, -3.8% to 6.2%. The problem is that you're much, much more vulnerable in those situations if the offense decides to air it out anyway. In those same situations, when the offense decides to throw a pass, base defense DVOA soars to 18.5%, while nickel rises to 3.4%. In short, if you're guessing run, you'd better not be wrong.
Last year, we posited teams might want to try 12 and 21 personnel more frequently, taking advantage of teams going lighter in nickel by bringing in a second tight end or fullback, because base defenses had superior DVOAs against both of those personnel groupings. That pattern has held for the past few years, but didn't last season, when two-back personnel struggled against nickel. I'm going to blame that on the teams responsible for trying it rather than the personnel packages, however. The Falcons led the league with 107 snaps of 21 against nickel, and any stat where the 2021 Falcons are the largest part of your sample size is not going to turn out well. Teams such as Indianapolis, Cleveland, and San Francisco still had plenty of success running 12 against nickel, so perhaps the strategy should be amended to "get bigger if you actually have quality players to use."
Only three teams (Seattle, Pittsburgh, and Arizona) even hit 50 snaps of base-against-11 as that strategy continues to die out. Putting a linebacker on a slot receiver is a great way to give up tons of passing yards! It should be noted that while the Steelers and Cardinals struggled in those situations, Seattle actually had a solid -15.9% defensive DVOA in base versus nickel, which moves us quite nicely into our team-specific tables.
Team By Team Frequencies
|Defensive Personnel Frequency, 2021|
The Seahawks led the league by using base 38% of the time. That's a new low-water mark for the NFL, although only just—the Jaguars were at 39% in 2020. It worked for Seattle—they had a better DVOA in base than in nickel, both against the run (-23.6% to 0.0%) and the pass (15.9% to 20.4%).
Eight teams were better in base than nickel against both the run and the pass, which doesn't stand out as particularly interesting. What is interesting is that all eight teams were in the NFC, including all four teams in the NFC West—the Cardinals, Rams, 49ers, and Seahawks were joined by the Vikings, Saints, Giants, and Football Team. Base defenses had a -8.7% DVOA against NFC West teams, the best in the league. They were at -3.9% against the NFC as a whole, compared to -1.6% in the AFC, and all four NFC divisions were in the negative.
The conference situation is mostly a "splits happen" scenario, I think, but offenses in the NFC West specifically being bad against base defenses is confusing. It should be noted that we're noting two separate but related things here -- the four NFC West teams are both good when in base defenses, and bad when playing against base defenses. There's an 11.7% DVOA gap between NFC West offenses against base and nickel; the NFC East at 7.5% is the only other division above 5%. A lot of this is the Rams, specifically, struggling against base, putting up a -42.4% offensive DVOA against base defenses, but all four teams were down when opposing teams loaded the box against them. Defensively, the Rams and 49ers were leading the way in base with defensive DVOAs in the -20% range, but the Seahawks and Cardinals were above average there, as well. I have no explanation for this whatsoever; there's not any real similarities between the offenses or defenses in the West specifically that would seem to lead to this result. Vince Verhei suggests that since the secondaries of all four teams were subpar last season that removing a nickel corner for a front seven player was a natural upgrade for all four teams. I suppose that's a possibility, but something about that doesn't quite ring right for me. I don't have a better explanation, so some of this is just "look at the weird stat!" coverage, but man, strange things are strange.
I'd dismiss it as random chance … except the same thing happened in 2020, with defenses putting up a -8.1% DVOA against the NFC West in base. That's as far back as it goes, so it may just be a weird split happening two years in a row. But it is weird as heck. Maybe Pete Carroll is on to something relying on base defense as much as he has.
The Cowboys and Bills become the sixth and seven teams in our data set, going back to 2011, to use base on fewer than 10% of defensive snaps. Buffalo should be even lower—26 of their 54 snaps in base came against the Patriots in the Week 13 weather game where New England only attempted three passes; no need for extra cornerbacks in that situation. And then 18 more snaps in base came against the Patriots in Week 16. That just leaves 10 snaps the rest of the season in base defense for Buffalo; it was almost entirely an anti-Bill Belichick plan. As Mac Jones was held to 139 yards passing in the non-wind-affected game, I'd say the strategy worked, but it's very odd to see a team change their entire defensive philosophy only for one opponent.
As for Dallas? You have one snap of base against the Buccaneers in Week 1, one against the Chargers in Week 2, four against the Panthers in Week 4, and six against the Patriots in Week 6. The Cowboys' true "base" was Dan Quinn's 4-2, using a bunch of athletic players in the secondary for speed. But it's not like Quinn's teams in Atlanta were using base this rarely, or that Dallas was an outlier for nickel or dime packages before Quinn arrived. This was a significant change in style from Quinn.
And it worked out for the better—the Bills and Cowboys were first and second in defensive DVOA last season, both overall and against the pass. Is that solely from sitting in nickel all day long? No, but there is a 0.52 correlation between base usage and overall defensive DVOA. If you have the horses to play in nickel and dime on a regular basis without getting absolutely destroyed on the ground, it's more likely than not a good idea to at least give it a shot. Unless, apparently, you're playing Bill Belichick or the NFC West.
Changes From 2020
The biggest change from 2020 to 2021 occurred in Green Bay, where the Packers dropped from using dime 50% of the time to just 24%. The Packers had used dime over half the time in each of the previous two seasons,and they ranked first in dime usage from 2017 to 2020. Joe Barry replacing Mike Pettine produced a predictable drop in dime packages, and it didn't hurt that De'Vondre Campbell made the Packers linebacker corps a positive for the first time in ages. And with Green Bay using a first-round pick on Quay Walker, I would expect the Packers to remain a nickel-focused team rather than a dime-focused one, even if they stay near the top of the league in dime percentage.
Dime changes in general were behind the biggest shifts between 2020 and 2021. The Patriots dropped from 47% to 22%, the Panthers from 41% to 16%, and the Texans from 21% to 0%—Houston didn't have a single snap in dime last season after playing 221 snaps in it in 2020. In all four cases, this was mostly just sticking an extra linebacker on the field to add nickel snaps, although Carolina saw a fairly significant increase in base percentage (5% to 21%) to go along with it. That's closer to what Carolina used before Matt Rhule and Phil Snow came in, and I think it's related to Jeremy Chinn being used more as a true safety last season as opposed to a hybrid safety/linebacker. Chinn's role ended up mostly replacing Tre Boston's from 2020, and Chinn's old role was filled more by Jermaine Carter, leading to an increase in base percentage.
Wrapping around to finish back with Buffalo, the Bills didn't just avoid base—they avoided all non-nickel packages. They had just 85 snaps all season outside of their nickel; every other team had at least 231 snaps in other packages. Buffalo managed nine entire games without ever leaving nickel and were mostly forced out of it by circumstance. We mentioned the Patriots-forced base defenses, but Buffalo also only had 26 snaps of dime, all against the Chiefs when Matt Milano was inactive with a hamstring injury. They were back in nickel for the playoff game on all but two snaps; it was more about Milano than any fear of Patrick Mahomes and the Kansas City offense.
This is one reason why I'm rooting for a Buffalo Bills versus Los Angeles Rams Super Bowl this season. It would tickle me to see the team which never leaves nickel play against the team that never leaves 11, just for the sake of novelty.
20 comments, Last at 12 Aug 2022, 6:01pm
#19 by IdahoPat // Aug 12, 2022 - 5:56pm
... the Patriots' use of a sixth offensive lineman. It worked in the first game because it forced the Bills to get Taron Johnson out of the game and use Tyrel Dodson, primarily a STs core-four player. And the weather certainly dictated such an attack for NE. But in the second game, the Bills had AJ Klein back, a strong-side, run-stuffing LB who's a liability in the pass game. Even though NE was throwing more in the rematch, they were tipping their hand anytime they reported Onwenu into the game.
#18 by theTDC // Aug 06, 2022 - 4:48pm
“There's an 11.7% DVOA gap between NFC West offenses against base and nickel; the NFC East at 7.5% is the only other division above 5%. A lot of this is the Rams, specifically, struggling against base, putting up a -42.4% offensive DVOA against base defenses, but all four teams were down when opposing teams loaded the box against them.”
Bit late to this one, but I think this is just an extreme example of situational football. McVay is a great coach, maybe the best in the league, but any Rams fan will tell you that he does have some obvious play calling issues. I can’t tell you how many times on second and short he’s gone for an obvious run play, only to get stuffed.
I mean, I get what he’s thinking. It’s second and short, so the defence should be expecting the deep shot, so he zigs when they expect him to zag and runs instead. Except that he never goes for said deep shot, so the teams just stack the box and he runs into it anyway. It’s mind boggling that someone so intelligent does things like this so often. It’s possible he was mindful of Stafford’s tendinitis in his throwing arm, because that tendency of his appeared to be worse this year. But I do remember this from previous years, so I’m not sure either way.
#16 by JacqueShellacque // Jul 31, 2022 - 8:31am
[I]t's very odd to see a team change their entire defensive philosophy only for one opponent.
Against an important opponent, don't you want to show them something they haven't seen? This doesn't sound odd at all.
Also even in nickel the Bills D is in the top half of the league, but certain opponents (Colts) have seen and exploited vulnerabilities there.
#17 by Mike B. In Va // Aug 01, 2022 - 8:12am
It will be interesting to see if Buffalo's one major weakness (run defense) was addressed by the changes to the D-Line. I expect a worse defense this year, both because of the White injury and the difficulty of remaining elite of defense vs. offense.
#20 by IdahoPat // Aug 12, 2022 - 6:01pm
Tremaine Edmunds missed the Indy game. It definitely showed.
AJ Klein missed the first NE game. It definitely showed.
Milano misses the KC game, and BUF plays more dime than it did in any game in 2020.
The Bills arguably have the best roster in the league, but if they're missing a LB they're going to really have to get out of the comfort zone on defense. And that's where they can be exploited.
#15 by andrew // Jul 30, 2022 - 4:32pm
Lining up in 4-5 against trips on first-and-10 is a bold decision, forced by injuries, but both the Vikings and Titans ran into the 4-5 on all five plays.
I don't know about TN, but to me this is an indictment of the Zimmer-mandated run-first vikings offense albeit in a small sample set.
#9 by Pat // Jul 29, 2022 - 2:33pm
Miami's a nominal "3-4" but I think the issue is that they tend to put guys who are nominally safeties in the box (or close to the line) instead of linebackers. Brandon Jones for instance had the most sacks of any safety last year, I think, and Next Gen Stats had a tweet where one game he lined up 72% of the time in the box. And the Dolphins blitzed their "safeties" more times than any other game in the past 5 years or something.
So in some sense that's almost a "is he really a safety if he rarely lines up like one" thing.
#10 by Aaron Brooks G… // Jul 29, 2022 - 2:53pm
That was a big trend in college 15-20 years ago -- moving a strong safety into the box to replace a LB and basically using them as a fast, undersized LB. The Giants used it too, but ran it more from of a nickel look.
#11 by Pat // Jul 29, 2022 - 3:51pm
It was much more extreme for the Dolphins last year, with the centerpiece being the Ravens game. You literally had plays that have 8 guys on the line (as rushers) and 3 in coverage. Except 3 of the guys on the line are technically defensive backs. (You can see some of these in a tweet breakdown from Baldinger).
So that's dime. Cover-0 dime. Belichick and Flores are nuts.
#4 by ImNewAroundThe… // Jul 29, 2022 - 10:40am
Top 1 in base = no playoffs. Bottom 3 = playoffs
Top 2 in nickel = playoffs. Bottom 1 = no playoffs
Top 1 in dime+ = no playoffs. Bottom 2 = also no playoffs
Conclusion: balance but lean coverage? Good DBs but at least one decent LB?
#5 by Pat // Jul 29, 2022 - 11:30am
Top 1 in base = no playoffs.
You mean top 2.
If you go out of nickel into base a lot, it means you're trailing and the other team is running out the clock. If you go out of nickel into dime a lot, it means you're leading and the other team is trying to catch up.
Correlate dime% minus base% vs win percentage, and the correlation's actually decently strong - R^2 of 0.13 - which might not sound like a lot but win percentage is noisy as hell intrinsically. (And nickel has no correlation, as you'd expect).
edit: put another way, the teams that ran dime more than base won 10.1 games on average, the teams that ran base more than dime won 7.8 games on average.
#2 by Travis // Jul 29, 2022 - 10:14am
Buffalo should be even lower—26 of their 54 snaps in base came against the Patriots in the Week 13 weather game where New England only attempted three passes; no need for extra cornerbacks in that situation. And then 18 more snaps in base came against the Patriots in Week 16. That just leaves 10 snaps the rest of the season in base defense for Buffalo; it was almost entirely an anti-Bill Belichick plan. As Mac Jones was held to 139 yards passing in the non-wind-affected game, I'd say the strategy worked, but it's very odd to see a team change their entire defensive philosophy only for one opponent.
Did the Bills do anything different in the Wild Card game?
#1 by Aaron Brooks G… // Jul 29, 2022 - 9:43am
That was as true in the 1920s as it is in the 2020s, as the history of defensive football is a fairly straight line pulling more and more bodies out of the box and into more versatile locations deeper on the field.
Cart before horse.
The history of defensive football is a fairly straight line of offenses pulling more and more bodies out of the box and into more versatile locations deeper on the field, and defenses mirroring it.
Which made a lot of sense prior to 1950, because they were the same guys. If your offense had an extra receiver, your defense was stuck with an extra DB. In a lot of ways, that made the trend more interesting, because opening your offense meant you also had to lighten your defense.
"Big" includes anything with three defensive backs and fewer than four defensive linemen (mostly 4-4-3 and 3-5-3)
Should be "four or fewer", right? Because a 4-4-3 has four defensive linemen.
In short, if you're guessing run, you'd better not be wrong.
The dime+ table suggests the converse is true, too. Don't be wrong about guessing pass.
If you have the horses to play in nickel and dime on a regular basis without getting absolutely destroyed on the ground, it's more likely than not a good idea to at least give it a shot. Unless, apparently, you're playing Bill Belichick or the NFC West.
I think the observation is that the Bills didn't have the horses to stop Belichick without getting destroyed on the ground. They ran base basically not at all against Indy and on about 1/3 of rushes in the first Pats game -- a game the Pats basically spent entirely in the single-wing. You can argue they only began to properly adjust in the second Pats game.
What form is Nickle taking? Are teams adding a safety or a corner? Are they removing linemen or LBs?
#6 by Pat // Jul 29, 2022 - 11:42am
The dime+ table suggests the converse is true, too. Don't be wrong about guessing pass.
Positive DVOA doesn't necessarily mean success. It just means better than average. If you're at 3rd and 20 and the other team gets 15 yards on a run, well, you'll just have to live with 4th and 5.
It's also partly true with the rush part as well: if teams are in base to prevent a team from running out the clock, if they gain extra yardage on a pass, that's a trade you're willing to take. But I think there are more cases of "4th and 1 pass for 15" there.
#12 by Pat // Jul 29, 2022 - 4:08pm
Even if they're successful plays it's not necessarily bad for a defense. Yes, yes, we all hate prevent defense, but it's not exactly insane. (Remember DVOA correlates with points not win percentage).
When you've got such a generic grouping as this ("dime" just describes who's on the field, not what a team is actually doing) it's tough to actually gauge what's due to the intrinsic factors and what's just situational.