The Offseason Skeptic Examines Quarterbacks

This offseason, I have broadened my continuing effort to help Washington fans avoid the heartbreak that comes from setting unrealistic expectations. In previous years, I have attempted to provide realistic benchmarks for the rookie class based on historical data from comparable players. In 2023, I have branched out to include returning players who have become the focus of seemingly over optimistic projections for the coming season.

The first three installments of the Offseason Skeptic examined the statistical reality of hopeful expectations developing around four players who have done little on the field to warrant the level of optimism being directed toward them. Second-year tight end, Cole Turner, failed to seize the opportunities presented by injuries to players ahead of him in the NFL’s fourth-weakest receiving tight end unit. Few tight ends in the past decade have carved out successful NFL careers after such a poor rookie debut. Fourth-year offensive lineman, Saahdiq Charles seems to have gained Ron Rivera’s confidence to enter training camp as the man to beat at left guard, despite failing to gain any traction on a unit that has struggled with injuries and losses of key starters throughout his first three years in the league. The coach seems to be counting on 2022 7^th round pick Chris Paul to hold man the position if Charles can’t. Chase Young enters his fourth NFL season facing long odds to prove that he can replicate his promising rookie season performance, let alone build on it.

The fourth installment takes a slightly different angle. Rather than asking whether 2022 fifth-round pick Sam Howell can live up to the loftiest expectations being set for him by some Hogs Haven readers and writers, this article will address another widespread expectation in the Washington fanbase, which sets us at odds with how the national media seem to view the Commanders’ young quarterback. That is, regardless of whether Sam Howell is the guy, long term, his debut season as the Commanders’ starter has to be better than 2022’s starting duo of Carson Wentz and Taylor Heinicke.

Photo by Rob Carr/Getty Images

I’m not even going to question whether the Commanders have found their future franchise quarterback. A few commenters on Hogs Haven have taken my occasional attempts to tap the breaks on the Sam Howell hype train to mean I am down on his chances. Nothing could actually be further from the truth. Sam Howell has been my guy since the 2022 draft, when I liked him as an option in the second round, if Rivera were to unwisely pass on Kenny Pickett (whom I assumed would be off the board in the top five) or Matt Corral, whom I thought would be a realistic possibility at their pick. I got so caught up in the excitement when he was drafted in the fifth round that I threw my usual skepticism to the wind and wrote a series of pre-season hype pieces (Weeks 2, 3, 4) which opened by comparing Howell to Russell Wilson and Dak Prescott entering their exceptional rookie seasons.

Sadly, Rivera chose to go another direction at quarterback, even sticking with his chosen veteran starter for the pivotal Cleveland game, despite later admitting that Howell was showing promising signs toward the end of his rookie season. Wentz, and backup, Taylor Heinicke, set a fairly low bar for Howell to clear. Neither achieved positive Expected Points Added per play on the season, meaning that when either QB had the ball, they tended to set the team back more often than they moved it forward.

Even so, I have to question whether it is realistic to expect a quarterback to be better than a pair of borderline NFL starters in his first season as a starter. The reason for my skepticism is that even the best quarterbacks can take longer than a season to adjust to the NFL. To illustrate what I mean, I’ll give you a choice of two quarterbacks to start your expansion franchise. Here are the stat lines from their first seasons as starters:

I think most readers will have picked QB B, who had a little higher passing productivity (Yards per Attempt), the same rate of scoring, and threw interceptions at less than a third the rate of QB A. In that case, they will have chosen to start Gardner Minshew over Hall of Famer Peyton Manning.

In his second season, Manning’s interception rate nearly halved and he made his first Pro Bowl appearance. The main point is that even the best QBs can have fairly pedestrian debuts as starters. Also, it is not unusual for QBs to throw a lot of picks in their first seasons.

How Much Can We Learn from One Game?

Compared to the other players featured in this series, we know relatively little about Howell in the NFL. All we really have to go on is a single start, in which he made 29 plays. To kick this off, let’s see how Sam Howell’s first start for Washington compared to each of the other QBs drafted by the team in the Snyder Era:

QBs’ first starts are ranked by Expected Points Added per Play (EPA/p). This is a composite performance metric, which uses a statistical model, drawn from over a decade of NFL data, to rate how much the QB increased (or decreased) the team’s chance of scoring, based on previous play outcomes in similar down, distance and game situations. You can think of it as a statistically valid version of PFF’s subjective player grades. EPA/p takes into account all the ways a player can contribute to, or detract from, their team’s chance of scoring. It is similar to ESPN’s QBR, which also weights plays relative to game situation to prevent dilution by players padding their stats in garbage time.

EPA rank is how the Washington QB ranked among NFL QBs playing the same week. Sam Howell debuted right around the middle of the pack. Of course, it was the final game of the season and many teams were resting their starters.

Howell’s first start in burgundy and gold ranks fourth by this metric, but he is pretty close to Jason Campbell in third place. Different metrics, such as QBR, put Howell ahead of Campbell. Howell had low total passing yardage, but he had the fewest passing attempts of any of the QBs. He had the second-highest passing productivity (Y/Att) and fourth-highest TD rate. He was also the most effective runner, overall, achieving slightly fewer yards per carry than Kirk Cousins, while also scoring on the ground and earning style points for levelling the defensive back to finish.

Neither player comes close to the stellar first starts by 2012 draftees Robert Griffin III and Kirk Cousins. Similarly, Howell and Campbell are well ahead of Patrick Ramsey’s first start in which he threw four interceptions and was sacked seven times. Dwayne Haskins is also well behind Howell and Campbell in EPA/p, but splits them in Passer Rating (Rate) and edges them out in QBR.

That doesn’t bode well for fans hoping that Howell will have a better career than the other QBs drafted in the Snyder era. Let me try something else.

Mark Tyler has recent set an extremely high expectation for Howell’s first season as starter, projecting a stat line which would have placed him in the top 10 of NFL QBs in 2022. Let’s see how Howell’s first start compared to those of the top QBs in the NFL.

To identify the top QBs of the last decade, I picked all the QBs who had ranked in the top 10 by QBR at least three times from 2013 to 2022. Peyton Manning only appeared twice, but I grandfathered him in because he led the league the previous season and was in the top three in six of the previous seven seasons.

The top QBs’ first NFL starts are ranked by EPA/p. In a few cases, I counted the player’s first NFL game playing the majority of NFL snaps. In most cases, that was the QB’s first game as designated starter. Peyton Manning’s rookie season predates the EPA calculator by one season, so I kind of slotted him in around where I thought he might fit.

Readers might get a shock by who’s on top, but the Amish Rifle had a pretty spectacular first outing for St Louis, in relief of injured starter Jamie Martin.

How did Howell’s first NFL start compare to this mostly illustrious group? Let’s have a look at the stats that can be compared across games with different numbers of plays. Howell’s 57.9% completion rate ranked a respectable 8^th out of 17. He was 5^th in yards per passing attempt, percentage of passes for touchdowns and Passer Rating (“Rate”).

On the not so good side he was tied with Lamar Jackson for the 5^th highest interception rate, as well as the 2^nd highest rate of sacks.

So, despite the relatively high interception and sack rates, Howell’s first game as a pro is not out of place among the first starts of some of the best QBs of the last decade.

Setting Unrealistic Expectations

What expectations would be reasonable to set for Howell in his first reason as a starter? Mark Tyler projected the following numbers:

16 games played

64% completion rate, 3980 yards, 25 TD, 12 Int

370 rushing yards, 4 rushing TD

Those figures would have ranked Howell 7^th among NFL quarterbacks in Total Offense, 8^th in TDs and rushing yards, and 10^th in passing yardage if he had played in 2022.

Just how good can Howell be? To get an idea of what the high end of the range of expectations might look like, I’ll start with a projection derived from the average of the first seasons of the top QBs of the last decade we looked at in the last section. Here are the stats from their first seasons as starters:

Since all of these QBs had their first starts in a 16 game season, I needed to scale everything up to 17 games. I started with games played (G) and games started (GS). The average QB in the table played 15 games and started 14 in his first season starting. Those figures equate to playing 94% of games and starting 89%. Multiplying by 17 games gives a projection of 16 games played and 15 started after rounding. I then calculated the number of passing attempts based on the per-game average, scaled up to 16 games played and used the average rate statistics (Completion rate, yards/completion, TD rate, INT rate, sack rate) to make the projection.

If Howell were to play like the average of the listed players in a 17 game season, he would achieve the following passing stats:

16 games played, 15 starts

283 completions, 468 attempts (60.5% completions), 3340 yards

21 touchdowns, 14 interceptions

84.7 passer rating

32 sacks, 197 yards

Therefore, if Howell has a first starting season like the average of the top QBs over the past decade, he will fall 640 yards short of Mark’s projection in total yardage, as well as falling short in completion rate, passing TDs and also throwing interceptions.

To project rushing figures for Howell, I limited the pool to the mobile QBs, since it wouldn’t make sense to base the projection on guys like Ben Roethlisberger and Peyton Manning who are grayed out. I estimated his rushing attempts based on the mobile QBs’ average attempts per game then used their average yards per attempt and TDs per attempt to project totals for 16 games. Here are his projected figures:

80 attempts, 388 yards, 4.9 yards per attempt, 5 TDs

These numbers are even higher than Mark’s. However, the rushing projection might not be unreasonable. He rushed for 1,106 net yards in his senior season at UNC, which was the third highest total by an FBS QB in the last three seasons.

As excited as I am about Howell, though, I can’t help feeling that we might be setting the expectations a little high for our second-year quarterback. If he’s not in the Pro Bowl by his third season, will people say he was a bust?

Reasons for Hope

Perhaps it might be a little unreasonable to expect Howell to debut like the best quarterbacks of the century so far. After all, very few quarterbacks selected as late in the draft, as he was, ever become NFL starters. In order to provide a more appropriate comparison, I did another projection based on players drafted after the second round from 2012 to 2022 who became starters in the first or second seasons. To keep this projection as close as possible to the previous one, my criterion for “starting” was starting seven or more games.

Here are the later round QBs’ first seasons as starters ranked by EPA per play:

Only 12 out of 81 (14.8%) QBs drafted after the second round from 2012 to 2022 became starters in their first or second seasons. So we are comparing Sam to a highly selected sample. However, only two out 12 (2.5% of 81 later round QBs) have become long-term starters. Those happen to be the two holdovers from the top QBs sample. Purdy could make it three. Only six out of 12 started for more than one season. Nick Foles has played like one of the best QBs in the NFL at times, but has never been able to maintain that consistently. Jacoby Brissett and probably Gardner Minshew have become high-end backups. The only other long-term starter drafted in the late rounds during this period, Kirk Cousins, missed out because he became a starter in his fourth season.

The later-round QBs tended to play in fewer games and have fewer starts in their first seasons at the helm than the top QBs we saw in the previous section. Interestingly, the later-round starters did better in a few areas including completion rate, percentage of passes thrown for TDs and interception rate. However, the top QBs as a group did better at performing when it counts, resulting in higher average QBR and EPA/play.

If anyone’s wondering why I picked an 11-year period instead of 10, it just seemed a shame to miss out on either Foles and Wilson or Purdy. Also, I included regular season and playoff games for Purdy to get the games started past threshold, but didn’t feel the need to do so for Wilson or Prescott since they started 16 regular season games, which is a sufficiently large sample.

Now for the fun part. I used a similar method as in the previous section to project season totals for a QB starting 16 games in a 17-game season to allow comparison with the combined performance of Carson Wentz and Taylor Heinicke in 2022.

The projection has Howell throwing fewer passes than Washington’s starters in 2022. That actually makes sense if Eric Bieniemy schemes to help his young QB out by leaning heavily on the run. I assume he drafted another running back for a reason. The projected completion rates and yards per attempt are slightly higher than Wentz and Heinicke’s, as is the percentage of passes thrown for TDs. As a result, the TD totals are close, despite Howell’s smaller number of passing attempts.

The biggest differences in the passing game between the projection and last season’s starters is in the negative plays. The projection anticipates a 28.6% reduction in interception rate, resulting in a 40% reduction in interceptions thrown. It also predicts an 11.5% reduction in the rate of sacks per dropback, resulting in a 26.7% reduction in sacks. The significance of a reduction in sacks is much greater than the loss of yardage, because sacks frequently kill drives, as we saw in 2022.

Sam Howell also demonstrated an excellent ability to improvise on broken plays and run the ball effectively in his first start. Heinicke also has that ability, but his wings were clipped in Scott Turner’s scheme in 2021 and 2022. The projection predicts a significant increase in rushing attempts and more than doubling of the rushing total from 2022, accompanied by a doubling of rushing touchdowns.

The net result of the increase in rushing production and decrease in yards lost on sacks is slightly lower total yardage production (-108 yards) and a modest increase in scoring (+1 TD) from the QB position. When combined with the significant reduction in interceptions and sacks, it is likely that the projected totals would represent a significant improvement in QB performance compared to 2022. Unfortunately, it is not really possible to project the elements of down and distance and game situation which are captured by EPA and QBR stats.

Photo by Jess Rapfogel/Getty Images

Conclusion

The fact that Sam Howell has earned the confidence of an NFL coaching staff to start in his second season means that he has already overcome long odds. Only 14.8% of QBs drafted after the second round since 2012 got that far.

If his first season as a starter is similar to that of the average later-round QB who became a starter in his first two seasons over the last 11 years, Commanders’ fans should notice an improvement in QB play compared to 2022. The model I used predicts a reduction in the number of passing plays, combined with improvements in passing efficiency, resulting in a modest drop in total passing yards and a small increase in touchdowns. This will be accompanied by a significant reduction in negative passing plays and big increase in rushing production.

Sam Howell is not likely to win Offensive Player of the Year if this projection proves accurate. However, if the offensive line holds up and the running game is at least as good as last season, Commanders can reasonably expect improved QB play to result in an increase in offensive production compared to 2022.

Acknowledgement: Edited by James Dorsett

---