Expected Points Added (EPA) is a broken terrible metric. I will now wait here for Ben Baldwin to come swooping into this article to yell at me. Any second now.
Ah, there he is. Good morning! So like I was saying EPA is terrible metric... when looking at punters. Now, I’m not going to dive into some argument about how EPA ignores context of a play formation or some other thing. No, EPA is flawed for punters because it takes into account two things that are absolutely out of the control of the punt team:
- The decision to punt in the first place.
- What the yards to go is on the punt attempt.
Let’s address the first point: Where on the field should you punt? We could write an entire article about that—and I plan to—but the decision is outside of the realm of what the punter or his coverage team can control. Not accounting for this can lead to some serious EPA penalties if we don’t eliminate it.
Should Michael Dickson receive negative EPA because Pete Carroll decides to punt from the opposition’s 45-yard line? Of course not, that hardly makes sense. If you wanted to grade the conservative nature of coaches? Sure, but leave poor Dickson out of these quibbles. Likewise, a decision to punt on 4th and 1—which you should almost never do—will punish the punter all over the field for this decision as seen below.
Alright, so now that I’ve convinced you EPA has some flaws in a specific regime, let’s fix them. The fix is actually pretty straight forward. We take every single punt from 2009-2017, lump them together by field position and yards to go, calculate the average for that yardage “bin,” and then we subtract this number from each individual punt in that bin. This will tell us at each specific distance and to go intersection if a punt was above or below average.
Now, for the disclaimers! Punting Expected Points Added Yards Adjusted per Punt (PEPAYA/P) extracts out a lot of information about coaches and offense, but it doesn’t remove the effect of having a bad coverage team. This metric is relatively stable year to year (.371 y/y), and correlates strongly to PFF’s grades for punters.
But we have a bit of a philosophical problem, because we’re about to dive into the realm of rare events. As you’ve seen me write about before, muffed punts are rare. So far in 2018, about 3% of all punts have been muffed, which is actually kind of high compared to some other years. But who do those muffs belong to? If we extract out muffed punts from our data set, we actually lose quite a bit of year over year stability, but it’s unclear that the punter should be given whole credit for these recoveries and drops.
To be precise, these are not counted in the same way as fumbles. They are distinctly flagged within the data, as they should be. Same goes for deflections and blocks. Thus far in 2018, there have been no blocks marked in nflscrapR, but two deflections, one of which hurts Michael Dickson’s PEPAYA/P numbers significantly.
If we want to look at the punting team as a whole, we should leave in fumbles, deflections, blocks—everything. This way we get a holistic measure of performance. If we want to judge the punter, we should extract out as much non-punter related things as possible, including fumbles, deflections and blocks.
The Punting Team
Not accounting for the strength of opposition and purely looking at our PEPAYA/P metric, the Seattle Seahawks are performing as the 7th best punting unit in the league. In general, the Seahawks are generating .25 EPA above average for every distance cross section of punt yardage when you look at the group holistically. That’s actually pretty darn good, and in fact only 20 teams are above zero using this metric.
But there is always room for improvement. So far, the punting unit has not generated a fumble, none of Dickson’s punts have been muffed—despite speculation that he would generate an inordinate amount of muffed receiving attempts— and Seattle is one of only two teams to allow a deflected punt. This occurred versus the Rams in Week 5, which actually hurts the team’s per punt numbers significantly and is frustrating, as it came on a blown blocking assignment by Shaquem Griffin and is only one of two deflections out of 803 attempts in 2018. But, since we are looking at the team, team mistakes are fair game.
Feel free to fight me in the comments—you all usually do but I still love you anyways—but I don’t think punters should be rewarded for fumbles or punished for blocks in the vast majority of cases. I do, however, think the punter is at least partially responsible for a muffed catch. But muffs come in many flavors, some are because of a fluttered ball and others because of a quick gunner, or maybe a team’s receiver plays the reception poorly.
Separating these out would require assigning at least a percentage of blame on each factor—an arduous task—and one I’m not entirely sure I’m qualified for. But, for the sake of argument, I am going to state that I believe the punter deserves at least some credit. It would also be a bit hypocritical of me to now insist that muffed punts shouldn’t count, after being so excited that Dickson would generate them. The irony is of course that if I exclude muffs, then Dickson is the best punter in the league by this metric. Without further equivocation, your top ten punters with deflections, blocks, and fumbles removed:
PEPAYA/P Midterm Grades: Punter A-, Punt Team B+
Dickson’s punts are generating a significant amount of expected points added over average for their yardage bins, even more so when we subtract out his one deflection—which was luckily mitigated by one Frank Clark. Thus far, talk that Dickson is not performing well are, to be polite, flat out false. It is true that Dickson could stand to improve, his one shanked punt is evidence of that. It can also be true that he’s not the best punter in the league, and thus far not even the best rookie punter—that would seem to go to Trevor Daniel of the Houston Texans, though PFF disagrees with this, even though in general PFF and PEPAYA/P agree strongly in most cases.
If ever there was a case to make that PEPAYA/P too heavily rewards muffed attempts, then Daniel very well could be it. They are, after all, rare events, and he’s had a muffed punt rate four times the league average this year. What will likely happen is that as the year progresses and our samples get larger, we will see PEPAYA/P and PFF converge even more strongly. In the meantime, looking around at all of our available metrics—PFF (11th), PEPAYA/P (3rd), special teams DVOA (8th)—it’s clear that Dickson is having quite the rookie season already, and his “true” ranking is likely somewhere in that range.
So, depending on your priors, Dickson is anywhere from the best to the 11th best punter in the league so far this year. Whether he was worth a fifth round pick (and a trade up) remains to be seen, but in my not so humble opinion he’s certainly on pace.
What is Michael Dickson’s punter rank this year?
This poll is closed