Tuesday, August 28, 2012
How to Use Football Stats: Punters
Most football fans don't pay much attention to punters. That's understandable. But among those who give any thought to the position, a majority simply assume that whoever has the highest punting average is the best. Even the league itself follows that model, making gross punting average the most prominent statistic at the position.
Unfortunately, gross average is an awful way to evaluate punters. Punting average doesn't account for hang time, directional kicking, or touchbacks. That means a 50-yard line drive that gets returned for a touchdown results in a better average than a 40-yard rainbow that draws a fair catch or immediate hit. A 50-yard bomb that lands in the end zone provides a better average than a 40-yarder downed inside the 10. I would even argue that gross punting average is a meaningless stat, completely useless except as a novelty.
Judging punters by where their kicks land, rather than where they end up after returns and touchbacks, is wildly misleading. In fact, punters on bad teams have better averages than punters on good teams.
Good teams, on average, are better at every position than bad teams. This is particularly true for specialists. A good punter is a luxury. General managers on bad teams don't look at their squads and say to themselves, "My goodness, we're terrible! Let's prioritize good punting, and maybe we'll finish 5-11 instead of 4-12." They look for a new quarterback, or offensive linemen, or cornerbacks. Good teams can afford to spend their money and draft picks on special teamers, and great GMs are the ones who get the little things right: the veteran guard who still has something left, the bargain linebacker, the overlooked punter or returner.
Since 2000, 55 teams have finished 12-4 or better, and 55 have finished 4-12 or worse. The great teams averaged 43.0 yards per punt, compared to 43.5 for the terrible teams. Thus, punters are unique, the only position in major professional sports in which the stat by which players are most often evaluated actually favors those who are worst at their jobs.
There are two main reasons punters on bad teams have better gross averages than those on good teams:
1) Clueless GMs on poorly-managed teams judge punters by gross average and don't recognize the importance of hang time or directional kicking.
2) If your offense stinks, you punt from your own territory most of the time, so you have the whole field to work with. Punters on bad teams can bomb away, as far as they can kick. Teams with more productive offense often punt from beyond midfield, so the punter has to shorten his kicks to keep them out of the end zone. There's no such thing as a 50-yard punt from the opponent's 45-yard line, and even a 45-yarder is a very poor play in that situation. A 35-yard punt looks bad in the stats, but it's a good play from the opponent's 45.
If your offense stinks, you'll punt from your own territory and get to use the whole field. If your team advances the ball, you have less room before you have to worry about the end zone.
So bad teams don't really have better punters than good teams. If you broke it down, you would find that even over small differences — say 9-7 teams rather than 8-8 — the better teams are better, on average, at every position. They have better quarterbacks, better running backs, better linebackers, better punters. Gross punting average not only says 8-8 teams are better than 9-7, it says 4-12 teams are better than 12-4. It's not merely a bad way to evaluate punters, it often yields the opposite of the truth.
Fortunately, we can account for factors like hang time and touchbacks by using net average rather than gross average. Net average looks at where the ball ended up, rather than where it landed, so it adjusts for return yardage and touchbacks. Problem solved? Well, let's be clear that net average represents a huge upgrade. Now you don't reward line-drive kickers, middle-of-the-end-zone bombers, and guys who think it's okay to punt to Devin Hester.
But net average isn't a great stat, either. Very good teams and very bad teams have comparable net averages, too, actually still a little better for the bad teams. So what are the problems with net average?
1. Punters on good teams frequently have to shorten up their kicks to stay out of the end zone. A 50-yard punt from midfield is not a good play, so a punter in that situation has a ceiling on his net average. Using net instead of gross stops rewarding touchbacks, but it still penalizes punters who have to work with a short field, and rewards those whose teams go three-and-out.
2. Good kickers and daring coaches help a punter's average. Why does the placekicker matter? Because if you have a great kicker, you try more field goals from 50 yards and beyond. A team that doesn't want to attempt a 54-yard field goal punts from the 36-yard line, which devastates a punter's averages, including net average. A 28-yard punt is a very good play in that situation, but it does horrifying things to one's average.
A team with a great placekicker might attempt a 60-yard field goal, which saves the punter from a short field. A conservative coach probably chooses to punt more often than to attempt the long field goal.
3. Bad weather has a significant impact on punters. This factor is sometimes overstated, but cold weather and strong wind can limit kicking range. It's not fair to expect a punter in Chicago to have the same numbers as one in San Diego. Of course, domed stadiums are even better than warm weather. And even better than domes: Denver.
4. Return yardage is affected by the coverage team. If those guys aren't any good, it hurts the punter's net. The coverage team is also responsible for downing punts that might roll into the end zone, and a punter needs blockers to give him time to kick. Those factors don't usually play very much role in a punter's stats, but they certainly aren't the same for everyone.
5. Not all yards are created equal. The most important yardage differences in football are those within 10 yards of the goal line. Pinning an opponent deep creates the possibility of a safety, and limits the offense's options, often earning an easy three-and-out. Skill in pinning opponents deep is arguably a good punter's most important attribute. Net average says a 40-yard punt from your own 20-yard line and a 40-yard punt from the opponent's 41-yard line are equally valuable plays, which obviously isn't true.
Because net average is an average, it doesn't account for big plays, like TD returns. Take two punters, Adam and Zeke, who each punt twice from their own 20. Adam's first kick has no hang time and gets returned for a touchdown. His second punt goes over the returner's head and bounces 20 yards downfield, finally rolling out of bounds at the other 20-yard line, a 60-yard punt. Both of Zeke's kicks get returned to his 40-yard line, a net of 20 yards per punt. Net average recognizes that Adam and Zeke have both had very bad days, but it regards them as equal. Adam's disastrous line drive returned for a TD only counts as a -20 net, since an average sees every yard as equally valuable.
So even net average isn't a great way to judge punters, because it's influenced by team factors: the quality of the blockers and coverage team, the field position of the offense, and even the quality of the placekicker and the philosophy of the coaching staff. Net average also doesn't account for weather conditions or quality of opposition, and it treats all yards as equal, with no recognition for punters who can hit the coffin corner and no aversion to those who give up big plays.
Every football player has the same fundamental goal: to help his team win. For punters, this is done by maximizing field position. Specifically, a good punter aims to create bad field position for opponents. He can do this several ways:
1) Maximizing gross punt yardage, by punting really far, so that the opponent's distance from the end zone is maximized.
2) Preventing returns by inducing fair catches or kicking the ball out of bounds.
3) Pinning opponents deep and avoiding touchbacks.
4) Avoiding blocks.
5) Contributing to turnovers by creating situations in which the returner is more likely to mishandle the punt.
The punter usually has minimal influence on the last two points, both of which are sufficiently rare that they are difficult to measure and subject to flukes, but the other three are common situations, all of which are easily measurable with common statistics. There are now statisticians doing advanced study of all NFL statistics, including punting, but even with the fairly basic numbers available at NFL.com, you can do a decent evaluation of punters.
My favorite stats for this are: net average (explained above), I-20:TB (ratio of punts downed inside the 20 [I-20] to touchbacks), return percentage (percentage of punts returned by the opponent), and short field percentage (how often the punter is close to the opponent's end zone).
I-10:TB is even better than I-20, but the league doesn't keep that statistic, so it takes a little extra work. Leaguewide, I-20:TB average is about 3.5 to 1, but in a given year, the best punters will be at least twice that. This stat indicates a punter's ability to pin opponents deep, while avoiding touchbacks that give the team critical space near its own end zone.
I actually reverse return percentage, using a stat I call not% — the percentage of punts that don't come back: fair catches, downed by the coverage team, or out of bounds. Basically, everything except returns and touchbacks. Effectively, not% measures how often the gross average and net average are the same — higher numbers are better.
League-wide, not% is only about 43%, since it accounts for touchbacks. However, a good punter can keep his not% over 50, meaning less than half his kicks come back from where they were landed. This stat measures a punter's ability to get the most out of his distance, kicking directionally or with excellent hang time, and minimizing the potential for big plays. It's especially important for punters on teams with below-average coverage units.
Short field percentage (SF%) is a basic calculation: (I-20 + TB) / punts. What percentage of the player's punts were touchbacks or downed inside the 20? This helps determine which punters were most often faced with a short field that limited their opportunities to kick for distance. Without considering this, you'll conclude that most of the best punters play for teams with bad offenses and weak-legged placekickers.
I'll run you through the basic selection process I use for choosing all-pro punters. I consider film study to be of limited use for this position, so I rely almost entirely on statistics. This is not exactly how I look at the position, but it's close, and it's straightforward:
1. Begin with every full-time punter in the NFL, usually just 30 or 31 because of injuries and benchings.
2. Cut the bottom quarter in net average.
3. Cut the bottom quarter in I-20:TB.
4. Look at the bottom quarter in not% and cut anyone who plays in a dome, a warm-weather stadium, or Denver.
5. Cut anyone who rates in the bottom half of both SF% and net average.
That should leave a group of about 10 punters. At this point, I'm considering a variety of factors. Is there someone who ranks near the bottom of the remaining group in three or more of these stats (net avg, I-20:TB, not%, SF%)? Someone who's near the middle but played in a dome? What about other stats, like fair catches, or return average and TDs? That generally gets me down to a group of 4-6. At this point, I calculate I-10:TB and consider additional factors.
The most important consideration when evaluating punters is the weakness of average, both gross and net. By way of example, consider the most celebrated punter of the last 30 years, Shane Lechler. He's a good punter, with a tremendously strong leg and historic averages. He is the all-time leader in punting average (47.6), has four of the top 10 seasons in history, and has led the NFL in gross average six times. That's great, but Lechler's game also includes weaknesses that largely cancel out his fantastic averages.
1. His net average is substantially lower than his gross average. If you were to take every punter in the league, and try to figure out whose gross punting average is least representative of the actual field position he gains, Shane Lechler would blow away the field. Since 2000, when he became Oakland's punter, the Raiders have a 9-yard difference between their gross average (47.5) and net average (38.6). Lechler's personal averages are 47.6 and 39.0, but the latter figure doesn't include blocks, which bring the difference to a full nine yards. No other team in the NFL has a gross-to-net difference of even eight yards (though the Texans are at 7.99). Lechler's gross average is the farthest from his net by a whole yard. Over the same 12 seasons, the Falcons' gross-to-net difference is just 4.3 yards, with a leaguewide average of 6.4.
The difference comes mostly from Lechler's unique propensity for punting the ball into the end zone, but he also seldom kicks out of bounds or forces fair catches.
2. Lechler is the worst punter in NFL history when it comes to the most important yards, near the opponent's goal line. He has led the NFL in touchbacks five times and has eight seasons of double-digit TBs. Over the course of his career, the rest of the AFC West has combined for just three seasons of double-digit TBs. Lechler averages 11.2 touchbacks per year. Everyone else is between 4-8; the next-worst team, Tampa Bay, averages 7.9 TB/yr.
14.3% of Lechler's punt become touchbacks. The next-worst rate over the last 12 seasons (the span of Lechler's career) is 10.6%, by the Titans. When it comes to bombing punts into the end zone, Lechler is all by himself. Per-season averages, 2000-11:
The NFL has recorded punter's touchbacks since 1991. Over that time, Shane Lechler has — by far — the worst ratio in history of punts down inside the 20 to touchbacks (2.25 to 1). Everyone else near the same level gets cut from the NFL after a year or two.
3. The Raiders have been terrible for most of Lechler's tenure. From 2000-11, the Raiders are 78-114 (.406), and Lechler's averages are in large part a product of how often he punts from his own territory. Oakland actually has the highest SF% in the NFL over these years (46.4%), but it's an illusion created by Lechler's ability to get touchbacks seemingly from anywhere on the field. The Raiders haven't been crossing midfield a whole lot in the last nine seasons; they rank 20th in scoring. Nor does the team ask Lechler to punt from close range: Raider placekicker Sebastian Janikowski has attempted more field goals of 50 yards or more (66) than any other kicker since 2000 (Jason Hanson is next, 53).
4. Lechler kicks mostly in favorable conditions. Oakland is not the easiest place in the NFL to be a punter, but it's certainly not the hardest. The average annual temperature in Oakland (59° F) is 11th-warmest of the 32 teams in the NFL, the stadium is not known for wind, and the Raiders also make annual trips to sunny San Diego and the thin air of Denver, which is every kicker's best friend.
5. Lechler allows a ton of returns. We know that punters don't just need to kick far, the ball needs to stay somewhere close to where it landed — long returns defeat the purpose of a deep punt. Lechler's not% is the worst of any contemporary punter, 34.5% (compared to a league average of 43.1%), and even omitting touchbacks, 60% of his punts are returned, 2nd-worst in the league since 2000 (the Lions top out at 61%; the league average is 52.6%). If you just go by live punts (no touchbacks, no out of bounds), 66% of Lechler's punts get returned. That's 32nd out of 32 teams.
Moreover, Lechler allows 11.2 yards per return, tied for the league-worst since 2000. That 11.2 return average is more than the lifetime averages of Joshua Cribbs (10.8), Dante Hall (10.5), and Eric Metcalf (9.8). Everyone who returns punts against Lechler becomes a superstar for the day. On average, Lechler yields 448 return yards per season, compared to a league average of 339 — over 100 more for Lechler.
I've heard Lechler fans defend him by claiming that the Raiders' special teams coverage unit isn't very good. That's probably true, but a good punter accounts for that with extra hang time (Lechler has the lowest fair catch percentage in the NFL) or by kicking away from the returner (8% of his punts go out of bounds, compared to a league average of 11%). I'm confident that Oakland's coverage teams have often been substandard during Lechler's tenure, but a talented punter can adjust for that, and Lechler continues to bomb 'em as deep as possible. That's the worst possible approach if your coverage unit is poor.
Essentially, Lechler's punts go too far and not high enough. Punt returners are, first and foremost, great open-field runners. These are guys who may not see much action as running backs or wide receivers or cornerbacks, but who can do amazing things when you put them in space. Really long kicks put returners in space. I'm not saying shorter kicks are better than longer kicks — all things being equal, they're obviously worse — but the last thing you want is to get an opposing returner in the open field. That's a recipe for long returns and enemy touchdowns. It sounds strange, but a 40-yard punt is often better than a 60-yard punt with equal hang time. If the coverage team has surrounded the return man before the ball gets there, he's obviously not going anywhere.
From 2000-11, Oakland ranks last, second to last, or tied for last in punt return yards allowed (5,379), punt return average allowed (11.2), fair catches (133), percentage of live punts returned (66%), percentage of punts where net yardage is worse than gross (65.5%), touchbacks (134), and inside-the-20 to touchback ratio (2.25 to 1). The most telling statistic, to me, actually combines two largely unrelated stats. As best I can tell, Lechler is the only regular punter in NFL history with more touchbacks (134) than fair catches (131). An average punter forces more than twice as many fair catches per season (16) as touchbacks (7). Lechler averages 11 of each.
Despite how it might seem, I'm not trying to bash Shane Lechler. He's wildly overrated, but he's not a bad punter. He's just as bad as a punter can realistically be and still lead the league in net average (which he has done four times). Lechler happens to be a perfect example of how most fans and journalists misinterpret punting stats, because he's great at the one thing they all look at (distance) and terrible at everything else (hang time, directional kicking, precision punting near the goal line), with built-in benefits playing for the Raiders (great placekicker and aggressive coaches, poor offense, nice weather).
Judging punters just by net average is like judging quarterbacks only by passer rating. There's a lot that's left out. Judging punters by gross average is like judging quarterbacks by rushing attempts. You can't get the whole picture from one stat. I realize that means analyzing punters takes more than three seconds, but being fair is more important than being fast.
I understand that most people don't care about punters and punting statistics. That's fine. But when people who don't know how to evaluate punters make bold proclamations or start talking about the Hall of Fame, that's irresponsible. It's shameful that every year a bunch of writers who are mostly concerned with getting quotes after the game select a Bay Area punter (Andy Lee or Lechler) with a good average and no finesse as All-Pro. It's misleading to fans who count on expert opinions, and it's unfair to smart punters like Brad Maynard and Michael Koenen, who don't get a ton of distance, but don't give up a lot of returns and mostly stay out of the end zone. But I guess it works out well for smart GMs.
* Gross punting average is a meaningless stat, almost entirely unrelated to helping the team.
* Net punting average is heavily influenced by field position, weather, quality of the special teams blockers and coverage unit, strength of opposition, and coaching philosophy.
* Avoiding touchbacks and returns is great, but you've got to have the leg, too. Average does matter — but it's not all-important.