With the draft upon us and most of free agency done, it’s time to look at the offensive identities of this year’s teams. Before I get to the 2019 coaches individually, what offenses they ran in their current and previous jobs, and what players will benefit or suffer from that identity, I want to take a look back at the identities of the 2018 teams.
I’m going to start with the most basic part of an offense’s identity, how many plays it runs from scrimmage:
The average team had 1007 plays from scrimmage in 2018 (down 8 plays from 2017); with a standard deviation (SD) of plus or minus 50 plays. Teams in green had at least one SD more plays than average; those in red were at least one SD below average.
Number of plays is somewhat connected with pace of play – the faster a team lines up and snaps the ball, the more plays it gets off. But it’s not just that. Incompletions stop the clock, so teams that pass more tend to have more plays too. Good offenses stay on the field more, also translating to more plays. Good defenses get the ball back, more plays on offense. Bad defenses mean a team trails, therefore passes more, therefore, more plays.
You can see that some “green” teams were good offenses, like NE, LAR, IND, and PIT. BAL was an example of a good defense and ball control/run-heavy offense leading to more plays. Last year, none of the really bad teams were even good enough to get a lot of plays off.
MIA ran just 878 plays from scrimmage, the 2nd lowest since the league went to 32 teams in 2002 – only SF 2005 had fewer (865). But MIA didn’t have the worst offense in the league – the Dolphins were 26th in points scored, tied for 27th in Pro Football Reference’s Offensive Simple Rating System (OSRS), and 26th in Football Outsiders’ Offensive DVOA. Bad, but not the worst. ARI, however, which also was 2 SDs below average in plays run, was perhaps the worst – last in all those categories.
Of the teams “only” 1 SD below average, CIN and TEN were mediocre to bad offenses, but the Chargers actually were quite good: 3rd in Offensive DVOA, 8th in OSRS, and 6th in scoring. So number of plays from scrimmage is not the only factor that goes into determining whether an offense is good or bad in reality.
Similarly, there is a lot more to fantasy scoring than just plays. The next chart shows how many total fantasy points (FP) each team accumulated rushing and receiving last year (10 yds = 1 FP, TD = 6 FP, receptions = 1 FP).
The chart shows the number of plays on the x-axis, with each labeled increment representing one standard deviation from average (1007). The y-axis depicts the total rush and receiving FP: the average team scored 1164, up quite a bit from 1083 in 2017 but in line with the 1152 FP in 2016. Again, each increment is one SD from average.
The dotted blue line represents the expected number of FP scored by a team, given a certain number of plays. The equation of the regression line is shown: Total FP = 1.852 * Number of plays – 701.97.
Teams above the dotted line scored more FP than would be expected by the number of plays they had: these were relatively efficient offenses. Those below the line were inefficient, scoring fewer points than they would be expected to.
KC led the league in total rushing and receiving FP but was below average in number of plays. There was a cluster of teams an SD above average in both FP and plays (PIT, LAR, IND, and TB) but also a couple of fantasy-productive offenses (ATL, NO) closer to KC in plays from scrimmage.
ARI (hiding down in the lower left of the graph) was easily the worst offense for fantasy (2 SD below average in FP) as well as for the “real” NFL.
MIA actually generated a few more FP than would be expected given its low number of plays – this is an example of how being efficient can still mean “bad, just not as bad as expected” since the Dolphins are on the same low level as several other teams in actual FP scored. Note that BUF had fewer FP than MIA despite 130 more plays from scrimmage. Four teams (TEN, WAS, NYJ, JAX) were strung between those two extremes in plays on roughly the same poor performance level of FP.
BAL, at the far left, had a huge number of plays but a little below average FP total. In fact, if you measure inefficiency as distance below the expected FP line (the blue dotted regression line), BAL was barely more efficient than BUF (the most inefficient by that standard). Of course, that is FANTASY inefficiency – the Ravens were not trying to help your fantasy team.
There were a total of six teams more than 100 FP below the regression line, i.e. very inefficient. Three, ARI, BUF, and NYJ, had rookie QBs taking most of the snaps. BAL had one taking just under half the snaps, and an ineffective veteran (Joe Flacco) taking most of the rest. The other two teams, JAX and WAS, had QB issues with their veterans, too.
The highly efficient teams (100+ FP above the line) generally had very good QB play: ATL, KC, LAC, LAR, NO, PIT, and TB. The Bucs stick out as an outlier from that group and some may question whether Jared Goff or Sean McVay deserves the credit for the Rams offense.
The R-squared number (0.347) tells us that # of plays and fantasy scoring are not that strongly correlated. Only about 34% of total rushing and receiving FP in 2018 can be explained by how many plays a team runs. Several factors work against the plays-to-FP relationship.
First, incomplete passes are a factor. They pull down the value of each play, although longer passes have a yards-gained benefit that somewhat offsets their lower completion percentage. But some passes essentially don’t have a chance of generating FP: spikes and throwaways, plus passes that are so inaccurate no target is assigned; all are wasted plays.
Second, kneeldowns, which have no fantasy value and can even hurt a QB’s rushing stats depending on how far back a QB kneels (or the spot/official scorer): of 394 kneeldowns last year,1 12 were ruled no gain (or loss), 19 were losses of two yards, 3 were for -3 yards, and 2 were for -4 (Sean Mannion vs. the 49ers in Week 17). The others were all ruled a loss of a yard, and from a fantasy perspective, were a wasted play.
A third type of wasted play is a sack, which generates no offensive FP (unless your league penalizes QBs for sacks).
If I throw out all passes with no assigned target, kneeldowns, and sacks, and call the remained “Total Fantasy Plays,” the R-squared of my regression line improves considerably, to 0.45:
The individual offenses don’t change a lot – although Buffalo moves quite a bit to the left. The Bills led the league in wasted pass attempts with 35 and had the 3rd most total wasted plays. The surprise leader in that category was Seattle, thanks to its high sack rate (51 total sacks), a lot of wasted pass attempts (27, 2nd only to Buffalo), and an above average number of kneeldowns (which is generally an indicator of a good team). MIA had the highest percentage of wasted plays, at 9.5% of its total plays from scrimmage. I need to do more research on how predictive wasted plays are from year-to-year, but generally, teams with fewer wasted plays are going to be more productive for fantasy.
Here are the wasted plays totals for last year:
WPA = Wasted Pass Attempts, i.e. those that have no assigned receiver as a target
KD = Kneeldowns
WP = Wasted Plays, or sacks + wasted pass attempts + kneeldowns
WP% = Wasted Plays Percentage; Wasted Plays divided by Total Plays from Scrimmage
Note that you want to be below average in Sacks and WPA and above average in KD. Only CHI, IND, LAC, NO, and PIT met those criteria; the Bears are the surprise on that list. Teams that were bad in all three categories were ARI, BUF, DET, GB, and WAS.
As I write this the day before the draft, I’m struck by how bad OAK looks in sacks and WPA, making me think the Raiders might be in the market for a QB tomorrow.
Another macro-level factor you might think is tied to plays from scrimmage is wins. In fact, wins – by themselves – explain little about a team’s number of plays:
A quick look at the equation for the dotted regression line tells the story: # of Plays = 6.3282 * Wins + 956.31 with an R-squared of 0.1291. In English: every game a team wins tends to add about 6.3 plays to their bottom line, but even a team with no wins “should” have 956 plays from scrimmage (which was about -1 SD below average in 2018). In fact, five teams last year had fewer plays than that, including the Chargers, who won 12 times. Wins explain only about 13% of the total number of plays a team runs. Winning teams DO tend to generate fantasy production, just not from play volume:
Again looking at the regression equation, FP = 31.547 * Wins + 910.48 or each win means about 32 more FP rushing and receiving. Not a huge amount, but not trivial. The R-squared (0.3247) indicates that wins explain about 32% of total rushing and receiving FP. So it’s a factor but not an overwhelming one. I’d generalize about the teams with more than 10 wins: those above the regression line had good offenses (NE, KC, LAR, and NO), those below it had good defenses (HOU, CHI, LAC) although you could argue that the Rams and Chargers had both.
Summary:
- Total plays explain about a third of fantasy scoring (rushing and receiving FP).
- Wasted plays (sacks, passes with no clear target including spikes, and kneeldowns) are factors in why total plays don’t more strongly correlate with fantasy production.
- If wasted plays are eliminated from a team’s total plays, the remaining number explains 45% of fantasy scoring.
- Wins are not a strong factor in how many total plays a team runs; they explain about a third of rushing and receiving FP as well.
As you can deduce from my comments about QBs and offensive (in)efficiency, I think QB quality is the biggest factor in explaining overall fantasy production of a team.
Part II will address the running game part of identity.
1Thanks to profootballreference.com’s Play Index.
