Football fans are in for an early treat this weekend. The Los Angeles Chargers and Baltimore Ravens face off during the early slate of games on Sunday, and the matchup looks to be a great one. According to our quality and game importance metrics, this is the game of the week by a wide margin, scoring a 94 in quality and an 86 in importance. With an overall game score of 90, it’s the highest rated game of Week 6 — and perhaps the most interesting aspect of the matchup involves the quarterbacks.
Justin Herbert and Lamar Jackson have led their teams to the top of their respective divisions. Their squads have identical 4-1 records, and according to a slew of metrics, the quarterbacks have produced at similar levels — yet they’ve arrived at those results in wildly different ways.
First, let’s look at the similarities. Both the Chargers and the Ravens are coming off last-second victories in Week 5. Herbert and the Chargers recovered from a 27-13 third-quarter deficit against Cleveland by scoring touchdowns on five of their last six drives to beat the Browns 47-42. And Jackson led the Ravens to a thrilling 31-25 overtime victory on Monday night, passing for touchdowns on four consecutive drives against the Indianapolis Colts.
On the season, Herbert’s and Jackson’s stats in key statistical categories are close to each other: They have an identical completion percentage (67.1 percent), each has passed for over 1,500 yards (1,576 for Herbert, 1,519 for Jackson), and both are throwing off-target balls around 2 percentage points lower than league average. Both Herbert and Jackson surpass 87-point passing grades from Pro Football Focus (87.6 for Herbert, 87.2 for Jackson), and the QBs have almost the same NFL passer rating while under pressure (104.7 for Herbert, 104.4 for Jackson).
But the similarities end when we look at when each QB is generating their production. On late downs — third and fourth down — Herbert has been incredible and Jackson has been, well, something less than that. Herbert ranks No. 1 in the NFL in raw QBR on late downs with a rating of 94.6. As a team, the Chargers are converting their third-down attempts nearly half the time (48.5 percent), a rate that’s 7.9 percentage points above league average. On fourth-down attempts, the Chargers are nearly perfect (7 for 8). But perhaps most impressive in L.A.’s late-down success is that some of Herbert’s longest passes have come on third and long — the hardest moments in which to make those big time throws.
Meanwhile, Jackson’s raw QBR on third and fourth down is verging on terrible, clocking in at 26.5, good for just 26th in the league. Considering how important third-down conversions are to team success (they help extend drives, after all) and how close Herbert and Jackson’s overall stats are, it’s remarkable how poor Lamar and the Ravens are in this phase of the game. As a team, Baltimore is converting its third downs just 34.5 percent of the time, 27th in the NFL and 6.1 percentage points below league average.
Things flip entirely when we look at early downs. On first- and second-down plays, Lamar is ranked No. 1 in the league (76.5 raw QBR), while Herbert is ranked 19th (48.5 raw QBR). Baltimore converts to a new set of downs without ever having to run a third-down play over 40 percent of the time (40.7 percent), good for fifth in the NFL. The Chargers escape facing third down far less, moving the sticks just 29.3 percent of the time on early downs, 23rd in the league.
It’s early in the season, so statistical anomalies should be expected, but it’s weird that two teams and QBs with such similar top-level statistical profiles earned their production in such different situations. And the fact that Herbert is succeeding on late downs and struggling on early downs is even more strange.
An interesting question to ask is: Which type of team would you prefer? One that excels on third and fourth down by overcoming the odds and extending drives, or one that earns its yards on early downs and avoids third downs entirely?
One way to try to answer this question is to look at the stability of team expected points added (EPA) per play on early downs and compare it to late down EPA per play over the course of a season. To do this, we calculated how well the year-to-date EPA per play of a team predicted the EPA per play of the same team in the upcoming week. We did this for each week of the regular season (starting Week 2) from 2000 to 2020, and for both early and late downs.
The results of the analysis indicate that first and second-down performance tends to be more predictive of future performance over the course of the season.