Last month, I had a mini tweet storm about punting value versus punting skill, where I discussed punters, how to measure punting, and how to reconcile the apparent gap between value (as measure by expected points) and perceived skill. Then I forgot all about it and never brought it up again. Today, I want to revisit some of my ideas, expand on them, and offer them for debate among the smart readers of this site. I know punting isn’t the most exciting topic, even for die hard football fans, but I used to punt in Pop Warner and have always been a bit of a punting aficionado. If you don’t care about the topic, come back later for something different. If you enjoy the subject, then strap in.
First, I gathered all punts from 2009-2017. Then, I removed those that came from non-punters. Next, I excluded blocked kicks, scoring plays, and turnovers, which are rare events that dramatically skew data and may not actually be indicative of a punter’s ability. I was left with 18375 punts ripe for examination. 1
To try to assign expected value to punts from a given yard line, I looked at the Expected Points Added by all punts in the data set.2 I then plotted yards from end zone (e.g., kicking from your own 20 is 80 yards from the end zone) against EPA and used a sextic function to find a best fit line for what our expectations should be for a punt from a particular yard line.3
Punting Value Versus Punting Skill: In Theory
I tend to think of punters in two categories: coffin corner punters and distance punters. Coffin corner punters are known as technicians for their incredible ability to spot the ball seemingly wherever they want to. They are pretty clearly the more skilled punters. Distance punters often put up big kicks because they play on bad teams that punt often from deep inside their own territory. They are seen as the brute force, unskilled bangers of the punting community. There are also guys like Johnny Hekker, who can do it all with aplomb, but special cases are exactly that – special. For most of my life, I have been on the side of the technicians and derided the big-legged guys who seem to lack control over their kicks. However, evidence suggests that, while the booming kicks may take less finesse, they may contribute more toward winning games.
Coffin corner kicks require a deft foot, but they also tend to produce less value from an EPA perspective. This is because the kicks often come from midfield or opponent territory, where EPA often advocates trying for a conversion or field goal. Punting sacrifices possession of the ball for field position, and because it is similar to a turnover, it is difficult to achieve a high EPA on any punt. This is especially true when teams opt to cede possession of the ball at the expense of a scoring opportunity. The red trend line on the chart below represents the expected point value of a punt from the corresponding yard line. Notice that the red line drops below zero inside of a team’s own 10 yard line and once a team reaches its own 45 yard line. This means EPA sees the idea of punting from within 55 yards from goal as an automatic negative, making it impossible for even the best punt to produce positive EPA in that situation.4
It’s not that the punters aren’t good or aren’t valuable, it is that they are being asked to do a job that, even when executed perfectly, is a negative EPA proposition. Here’s an example: in a 2014 game between the Cardinals and Lions, Bruce Arians called for Drew Butler to punt from the Detroit 36. The result of the play saw the Lions take possession on their own 1 yard line. Aside from a fluke turnover, this was the best possible outcome given the play that was called. For his trouble, Butler earned -0.360 EPA on the play.5
The expected points metric doesn’t know about special circumstances during games. It doesn’t know if a team’s kicker is injured or if the weather has basically removed anything but chip shot field goals from the playbook. For example: in 2013, Washington and the Chiefs were playing in the snow. The Chiefs were already up 31-0 in the second quarter, so Andy Reid called for a pooch punt from the Washington 29 on 4th and 6. It resulted in a touchback and a mere 9 yard change in field position. However, a missed FG would have resulted in a -7 or -8 yard change in field position. Given the situation, it’s hard to say punting was a worse choice than a FG attempt in the snow from 46 yards out.
EPA also doesn’t know if the mindset of a team is different when leading big. Using the previous example, with Kansas City up big, it is reasonable to assume that Reid wasn’t as interested in maximizing EPA on that play as he was mitigating risk against an inferior opponent. Metrics based on EPA tend to favor aggression, so conservative calls may be undervalued in rare scenarios like Reid’s.
Punting Value Versus Punting Skill: A Table
The table below contains every punter with at least 50 qualifying punts since 2009 and is an attempt to convey the most valuable and most skilled punters in that time frame. Read it thus: Johnny Hekker has 425 qualifying punts and has turned that into 103.9 expected points added, good for 0.245 EPA/P. His punts have been worth 0.221 Relative EPA/P, giving him 94.1 Relative EPA.
I’ll start with the obvious: Hekker is incredible. He combines the power of Shane Lechler with the precision of prime Mike Scifres. To top it off, he’s also adept at carrying out fakes, adding another dimension to his game that many punters simply don’t possess. Only twenty punters have produced positive EPA at the career level (min. 50 punts). Hekker’s 103.9 EPA is more than double runner-up Marquette King‘s 41.1. He also leads all players in EPA/P above average and, of course, total added value (in a landslide). When it comes to value on a per-punt basis, there’s Johnny Punting and rookie Rigoberto Sanchez, then there’s everyone else.
This study cuts off half of Lechler’s career, including five first team all pro seasons, so it is fair to believe he’d rate much higher were the data available. I have been hard on him for most of his career, but this study suggests that he has provided plenty of value, even when adjusting for the spot of his kicks, and even though his game doesn’t appeal to purists. I’m sorry, Shane.
Be sure to note the punters who have a positive REPA/P but a negative EPA/P. Those guys are lemonade makers, doing the best they can with the lemons their coaches have given them. They perform better than average, even though average is skewed by a boatload of bad decisions. Although they are good at their jobs, they are being used incorrectly and aren’t actually providing much value to their teams, at least from an EPA standpoint.
Rare is the punter with positive EPA and negative REPA. This indicates that their coaches used them wisely, putting them in position to produce positive EPA plays, but the punters themselves weren’t anything special. Thus far, only Marquette King (2014) and Britton Colquitt (2016) have done this in a season (min. 20 punts). No one has done it at the career level.
A couple bits of trivia…
Best punt without a turnover or silly backwards return: Facing the Chargers in 2013, Cincinnati faced 4th and seven from their own 21 yard line. Kevin Huber helped the defense out in a big way when he boomed a punt to the San Diego 4. That 75 yard shift in field position was worth 3.544 EPA and 3.457 REPA. It wasn’t the deciding factor in the game, but it was an important component in setting up the Bengals’ subsequent scoring drive.6
Worst punt without a return or block: Panthers punter Brad Nortman punting 6 yards from the Carolina 32 to the Carolina 38 against the Bears in 2012. Seven plays later, Jay Cutler would find Kellen Davis over the middle for a touchdown. Carolina ultimately lost 22-23. That punt was bad. Very bad. It was worth -3.077 EPA and -3.288 REPA.
One thing that is evident from the good and bad punts highlighted above is that, occasionally, a single punt can have a significant impact on the outcome of the football game. Most punts aren’t quite that important, but it is worth observing the cumulative value of all of those seemingly insignificant plays. Because the data set goes back only to 2009, we can’t get a clear picture of Tommy Davis taming the wild winds of Kezar Stadium, Ray Guy testing the adynaton “when pigs fly,” or unheralded Scifres hitting the coffin corner of the coffin corner. However, we do have the ability to examine the current and future crop of optional turnover specialists. That’s pretty exciting, at least for those who care about punting.
- Ideally, we could include hang time in the equation to see if that proves to be a significant factor. Anecdotally, a team would prefer a 40 yard punt that spent 4.3 seconds in the air to a punt of equal distance that took only 3.6 seconds to land. It would be nice to test to see if that anecdote is founded in reality and not just a vague idea of “common sense,” or whatever term people who prefer not to think too much call it. Unfortunately, I have only seen the Rams and Raiders consistently mention hang time in their game logs. ↩
- For EPA, I am using Ron Yurko‘s version. ↩
- The function, for those interested: y = -0.000000000743805x6 + 0.000000263921778x5 – 0.000037400877394x4 + 0.00269765455x3 – 0.105061671689785x2 + 2.1744406183205x – 20.578738996142. Obviously, there’s no real reason to show this many significant digits, but I wanted to be sure to show clear differentiation in the small numbers. This function produced an R2 of 0.13. ↩
- With the exception of a muffed punt, but I haven’t seen any credible evidence suggesting forcing turnovers is a real punting skill and not just a product of chance. Also note that going for it from midfield can also be a negative EPA proposition, if the distance is long enough. In those cases, a team is left with picking the “less bad” option. So it’s not “never punt from inside the 50” as much as it is “think twice about punting from inside the 50.” In general, avoid nevers and alwayses. In the future, I plan to revisit this specific aspect of using EPA to judge punting and the coaching decision to punt. ↩
- He would have earned a much worse number had Justin Bethel not made a great play to prevent a touchback. He would have earned a much, much worse number had the officials correctly ruled to allow Jeremy Ross‘s return. ↩
- The Chargers, deep in their own territory, failed to move the ball and punted from their own 14. Cincy took possession at their own 33 and scored a touchdown in ten plays. The Bengals won the game 17-10. ↩