DVOA Analysis
Football Outsiders' revolutionary metrics that break down every single play of the NFL season

# The Special Teams Manifesto

by Aaron Schatz

When I first created the VOA (Value Over Average) rating system over the offseason, it only included offense and defense.  I finally had time to develop special teams at the beginning of this season, and the special teams rating was introduced into the VOA ratings back in Week 6.  I also wrote that "a longer article on the special teams ratings is coming in the future."  I think I've promised this article more often than the Democrats have debated over the past three months, and that is a lot of debates.  This thing has been in gestation so long that Michael David Smith began to refer to it as "The Special Teams Manifesto."  Here it is, finally, accompanied by 2003 special teams statistics.

Before I begin to describe the different methods used for various parts of special teams, I should point out a couple of things.  The data here is my unofficial 2003 data.  Some of it still needs to be cleaned up so there are a few tiny play-by-play errors, a yard or two here or there.  The "league averages" used for the purposes of comparison are the 2002 numbers; one part of the offseason will be putting the 2002 and 2003 numbers together to determine new averages.  (The exception here is field goals, where I recomputed new "league averages" after Week 9 of 2003 that included a year and a half of stats instead of just one year.)  The numbers here may differ from data seen on the site previously because I've already started doing some of that aforementioned cleanup.

Let me warn you, there's some math ahead, so be prepared.

## KICKOFFS

Kickoff ratings compare each play to the league average based on point value of field position at the position of kick, catch, and return.  There are two parts of the kickoff play, of course: the kick, and the return.  I've determined a league average for how far a kick goes based on the yard line from where the kickoff occurs (almost always the 30-yard line) and a league average for how far a return goes based on the yard line where the ball is caught.

For the kicking team, the rating is based on net points compared to average, taking into account both the kick and the return if there is one.  For the return team, the rating is only based on how many points the return is worth compared to average, based on the location of the catch.  The return team has pretty much zero impact on the length of a kick, after all, so it makes no rate kickoff returns based on net kickoff yards rather than just return yards.  The return team's rating also does not consider kicks that don't have returns.  That includes touchbacks, kicks that go out of bounds, and some kicks that just don't get returned (note that a kick without a return is different from a kick with a zero yard return, which does count for the return team).  Touchbacks count as if they were kicks that landed on the 20-yard line with no return; out of bounds kicks count as if they were kicks that landed on the 40-yard line with no return.  It seems silly to give a kicker credit for a 71 or 75 yard kick if a kickoff goes into the end zone and the offense starts the next set of downs at their own 20-yard line.

Let me give an example using a single kickoff.  At the beginning of the third quarter of the Week 8 game between Kansas City and Buffalo, Rian Lindell kicked off to Dante Hall.  The kick went 59 yards to the Kansas City 11-yard line.  Now, in general the "next score" from your own 11-yard line will be -.70 points (that means the team without the ball is more likely to score first), and the value of the "next score" from the average catch location of a standard kickoff is -.85 points, so this kick is worth -.15 points to Buffalo.  Hall returned this kick to the Chiefs 33-yard line.  When the offense has the ball on their own 33-yard line, the point value of that field position is .76 in favor of the offense.  Since the return started with the point value at -.70, the total value of the kickoff return is 1.46 points.  The average value of a kickoff return from the 11-yard line is 1.36 points, so the value of the kickoff return is .10 for Kansas City, or -.10 for Buffalo.  Since the kick is worth -.15 to Buffalo, and the return is worth -.10 to Buffalo, the net value of the kickoff return is -.25 points below average.

The best type of return, of course, is one for a touchdown, usually worth about six and a half points to the return team.  The worst type of return, and best type of kickoff, is one where the return team fumbles the ball and loses possession to the kicking team, usually worth around -4 points to the return team.

Here are the best five teams and worst five teams in 2003 for net kickoff value, including both the kick and the return:

 BEST NET KICKOFFS WORST NET KICKOFFS ATL +21.3 SDG -18.5 PHI +16.2 JAC -17.7 DEN +16.1 CHI -15.3 STL +10.9 GNB -12.0 HOU +10.6 OAK -11.9

What happens when you separate the kickoff into its component parts, the kick and the return?  You'll notice that top kickoff teams are all different from the top five teams at stopping kickoff returns.  In fact, the best team at preventing kickoff returns -- Tennessee -- was also the worst team at gaining field position through the kickoff.  Counting touchbacks as kickoffs caught at the 20-yard line, and out of bounds as kickoffs caught at the 40-yard line, the NFL's average kickoff was caught at the 10-yard line.  The average Tennessee kickoff was caught at the 14-yard line, a yard and a half shorter than any other team.  In total, these short kickoffs cost the Titans 25.7 points in field position.  But the Titans made 14.3 points of that back by preventing good returns.

 BEST KICKOFF KICKS WORST KICKOFF KICKS BEST KICK RETURNS ALLOWED WORST KICK RETURNS ALLOWED STL +18.1 TEN -25.7 TEN +14.3 OAK -14.2 DEN +12.0 GNB -15.3 ATL +12.7 DET -8.1 NOR +11.3 JAC -13.2 PHI +8.9 STL -7.1 HOU +9.3 SDG -11.6 NYG +7.5 SDG -6.9 NWE +9.0 CHI -11.6 MIA +6.8 NOR -6.3

What about the other side of the kickoff?  Here were the best and worst kickoff return teams.  Touchbacks and out of bounds kicks are not included in this statistic.

 BEST KICKOFF RETURNS WORST KICKOFF RETURNS CHI +17.3 ATL -23.2 KAN +16.1 NYG -15.2 NYJ +15.4 STL -13.5 NWE +12.6 TEN -12.4 GNB +9.0 CIN -12.1

While most teams have the same player kicking the ball off all year, there are multiple players making kickoff returns, and a ranking of the top kickoff returners will look different from a ranking of the top return teams.  Here are the top five and bottom five kickoff returners for 2003, based on points over average.  Nope, Dante Hall is not on top:

 BEST KICKOFF RETURNERS WORST KICKOFF RETURNERS Azumah, Jerry CHI +21.2 Mitchell, Brian NYG -12.5 Hall, Dante KAN +18.0 Rossum, Allen ATL -11.1 Johnson, Bethel NWE +12.5 Berlin, Eddie TEN -9.0 Carter, Jonathan NYJ +10.0 Jervey, Travis ATL -6.6 Davenport, Najeh GNB +10.0 Morey, Sean PHI -6.4

While Dante Hall got all the press this year, Jerry Azumah was actually the best kickoff returner in the league, over the whole season (Hall's worth faded in the second half).  New Hampshire in the hizz-ouse.  Over on the other side, yes, that is Brian "holder of NFL kick return career records" Mitchell in last place.  Welcome to the New York Giants, where special teams go to die.  We're all very disappointed to see Sean Morey on this list.  We are ever true to Brown but we've got to tell the truth about the only NFL player from the alma mater of most of the Outsiders.  At least we're not responsible for Eddie Berlin, who was the player whose two fumbles cost the Tennessee Titans their important game with Indianapolis in Week 11.  Take out his five returns and the Titans are close to league average.

The fact that our kickoff return statistic does not take into account the length of the kick, except as a basis for judging the baseline average return, means that the quality of kickoffs against a team makes up a "hidden indicator" that can help explain past wins and losses.  Here are the teams that faced the best or worst combined kickoff quality from the opposition before their returners could even touch the ball.

 GAINED FIELD POSITION DUE TO BAD OPPONENT KICKS LOST FIELD POSITION DUE TO GOOD KICKS AGAINST BAL +15.2 ATL -14.8 WAS +10.3 MIN -8.7 SEA +9.8 STL -8.5 NYJ +9.4 DEN -8.3 PIT +8.4 NOR -6.7

Gee, do you think that stadiums may have some impact on the distances of kickoffs?  Four of the top five teams are dome teams, the fifth is thin-air Denver, and teams six and seven in kickoff value against are Houston and Indianapolis.

## PUNTS

Punt ratings work very similarly to kickoff ratings.  The difference is that I made the assumption that, unlike with kickoffs, the punt return team has some effect on the length of the punt itself.  Sometimes they block it, but more often they force the punter into an imperfect kick because he needs to get the ball off before a block can occur.  Kickoff ratings counted the net value of the kickoff for the kicking team, but counted only the return for the returning team.  Punt ratings, however, count the net point value of the punt for both the punting team and the return team.

Of course, we can separate out the two to see which teams are best at punt length (or preventing punt length) and which are best at punt returns (or preventing punt returns) despite the length of the punt.

Let's take a look first at the best net punting teams in the league, along with the main punter.

 BEST NET PUNTING WORST NET PUNTING BAL (Zastudil) +15.0 DEN (Knorr) -19.6 NOR (Berger) +13.6 MIN (E. Johnson) -19.6 OAK (Lechler) +11.2 NYJ (Stryzinski) -19.4 TEN (Hentrich) +11.0 DET (Harris) -17.7 HOU (Stanley) +9.9 ARI (Player) -17.1

Now, that includes blocked punts.  I'm not sure, if I want to list the best punters and punt return-prevention units, what to do about the problem of blocked punts.  How much responsibility does the punter himself have for getting his punt blocked?  And I would think that we would want to take blocked punts out if we wanted to see how well each team prevented runbacks.  So, here's a look at the best and worst punting teams in the league, punts only, blocks removed.  The main punter is listed with each team:

 BEST PUNT DISTANCE WORST PUNT DISTANCE OAK (Lechler) +23.1 NWE (Walter) -20.1 BUF (Moorman) +16.4 DET (Harris) -17.3 NOR (Berger) +15.5 DAL (Gowin) -16.8 TAM (Tupa) +15.1 MIN (E. Johnson) -13.1 CAR (Sauerbrun) +11.8 MIA (Turk) -12.5

How about preventing punt returns.  We should leave out touchbacks, right?  And let's leave out blocked punts, since punt blocking is a different skill than punt return prevention.  But what about fair catches?  Do punt returners make a fair catch because of a good punt team that is sitting on top of them?  Or only because of a good hang time, the punter's responsibility instead of the punt coverage team's responsibility?  What about when the punt team downs the punt without a return?

I'm honestly not sure what to include here, so I'm open to suggestions.  In the meantime, here's a list that includes everything but touchbacks and blocks, representing the five best and worst punt coverage teams in the league:

 BEST PREVENTION OF PUNT RETURNS WORST PREVENTION OF PUNT RETURNS MIA +19.5 DEN -18.6 BAL +19.4 OAK -15.1 DAL +17.0 STL -10.8 SEA +17.0 ARI -10.0 NWE +16.6 BUF -8.8

Yes, I do notice that many of the teams listed among the worst punting teams are also listed among the best coverage teams.  How to measure the importance of hang time vs. distance for punters?  This harkens back to Mike Smith's article defending Patriots punter Ken Walter.  However, of the top five punting teams, only one is also among the five worst punt coverage teams, Oakland.

Now for the other side.  Here are the best punt return teams in the league, with blocks and touchbacks removed.  Out of bounds punts and downed punts are both included, as are fair catches, since I need to do another run through the stats and the NFL.com logs to separate various kinds of "no return" punt returns (touchbacks are easy to separate out since they all take place at the 20-yard line).

 BEST PUNT RETURNS WORST PUNT RETURNS ATL +19.4 BUF -14.4 DET +17.9 WAS -14.3 KAN +14.9 ARI -14.2 OAK +12.9 TAM -13.8 PIT +12.8 STL -13.0

And here are the best and worst individual punt returners -- once again, this does include fair catches.

 BEST PUNT RETURNERS WORST PUNT RETURNERS Hall, Dante KAN +25.6 Lewis, Michael NOR -15.4 Randle El, Antwaan PIT +13.0 Brightful, Lamont BAL -15.0 Buchanon, Phillip OAK +12.9 Groce, DeJuan STL -14.3 Moss, Santana NYJ +12.5 Northcutt, Dennis CLE -12.0 McQuarters, R.W. CHI +12.2 Moses, J.J. HOU -11.9

Well, that's where Dante Hall's notoriety came from.  Of course, Hall did return two punts for touchdowns, and his rating is higher than you would otherwise expect.  Teams stopped kicking to him in the second half of the season, so he couldn't lose value.

Right now, the baseline average used to compare all punts and returns is the line of scrimmage for the punt itself.  In the offseason, I hope to re-do the punt return stats so it uses two variables for the baseline, both the line of scrimmage of the punt and the location of the catch.

## FIELD GOALS AND EXTRA POINTS

Field goals and extra points are judged somewhat differently from the rest of special teams.  For each distance of a field goal, I determined the average number of points per field goal attempt.  This created a jagged line; I then smoothed that out with the Excel trendline feature.  This chart shows you the actual numbers (blue) as well as the line used in the VOA ratings (red).  Once again, I'll note that field goals are the only part of these numbers where the baseline is based on more than just 2002 -- I redid these numbers after Week 9 of 2003.

Numbers for field goals of 54 yards on up are based on a limited sample , so those league averages are a bit more guesswork, and after about 60 yards I altered the equation to reflect that these kicks are very difficult to hit.  In fact, there was only one attempt over 60 yards during the year and a half analyzed to create this baseline.

 Neil Rackers, Week 16 Yds Pts NFL Avg Pts+ 49 3 1.75 +1.25 38 0 2.40 -2.40 33 0 2.59 -2.59 XP 1 0.98 +0.02 Total 4 7.72 -3.72

To determine how many points a field goal kicker is worth compared to average, just take the total number of points for that kicker, and subtract the league averaged based on the specific distances from which the kicker made his attempts.  To give an example, let's look at Neil "Patron Saint of the Loser League" Neil Rackers in Week 16, kicking for Arizona against Seattle.  Rackers kicked three attempts, only one of which actually went through the uprights.  In proper confusing Rackers style, it was the hardest one.  Rackers also attempted, and hit, one extra point.  When all was said and done, Rackers was worth 3.72 points less than an average field goal kicker attempting kicks from 49, 38, and 33 yards as well as an extra point.

I should note that after I came up with this statistic for field goal kickers, I went back and reread Pro Football Prospectus and discovered that they use virtually the same metric.  They don't count extra points, under the assumption that a missed extra point is random luck, but for the most part it is the same method, so I did want to give them credit for thinking of it first.  I do think counting extra points makes sense, partly because it really isn't any more work, partly because a missed extra point is really no more random than a missed 19-yard field goal.

 BEST FIELD GOAL KICKERS WORST FIELD GOAL KICKERS Vanderjagt, Mike IND +19.1 Marler, Seth JAC -18.3 Wilkins, Jeff STL +17.2 Pochman, Owen SFO -11.3 Hanson, Jason DET +13.0 Reed, Jeff PIT -10.1 Graham, Shayne CAR +9.7 Vinatieri, Adam NEW -10.0 Janikowski, Sebastian OAK +9.6 Lindell, Rian BUF -7.9

Mike Vanderjagt had a perfect season, as we all know, and Seth Marler was a loser league mainstay thanks to all his misses, but this list brings up one of the difficulties of football stats.  The limited sample can result in wild swings from year to year.  Last year, Adam Vinatieri was in first place, worth +12.7 points more than the average field goal kicker, but this year he has slumped horribly.  Meanwhile, Vanderjagt, Wilkins, and Graham were all worth about zero last year.  How many seasons need to be combined to get a sense of the true worth of a field goal kicker (or, for that matter, any football player)?

For every team that kicks a field goal, of course, there's a team that has a field goal kicked against it.  Field goals kicked against a defense are sort of a "hidden statistic" according to both our ratings and the statistics used by the NFL.  When do you ever an announcer say "Well, field goal kickers are good on 80 percent of attempts against the Rams" or anything along those lines?  It's possible, sure, that the defense has a small effect on whether a field goal kicker gets it through the goalposts.  A blocked field goal shows that the defense definitely had an effect on the miss, but with only one of these per week it doesn't seem like it would be easy to separate ability in blocking field goals from random chance.  In general, I think it is a safe assumption that the kicker has most of the responsibility for whether a field goal counts for three or zero.

However, while I don't think field goals kicked against a team say anything about that team's future performance, they are just as important as field goals kicked for the team when it comes to explaining past wins and losses.  Here are the top five and bottom five teams this year in this hidden indicator, field goals kicked against:

 OPPONENTS WITH WORST FIELD GOAL KICKING OPPONENTS WITH BEST FIELD GOAL KICKING WAS -14.7 BAL +12.8 CAR -10.2 GNB +11.6 SFO -8.3 SDG +8.9 NWE -7.9 DEN +7.6 PHI -6.2 OAK +7.3

Note that the team that received the most benefit from poor field goal kicking by opponents, Washington, did not block a single field goal this year.

## OTHER SPECIAL TEAMS

Two-point conversions and onside kicks are not included in these special teams ratings.  I'm not sure if there are enough onside kicks in a season for any difference between teams to be more than just luck.   I honestly wasn't sure what to do about two-point conversions -- count them as special teams or as regular offense/defense plays from the two-yard line? -- and I never quite decided.  In addition, nearly all fake field goals and punts are not included in the special teams rankings because they are listed as passes or rushes in the play-by-play logs.

## TURNING POINTS INTO VOA

Once I've totaled how many points above or below average can be attributed to special teams, another formula then transforms these numbers from points to VOA so the ratings can be added to offense and defense.  The estimate of how many points are worth one percent in VOA is probably off by a little bit and will be recomputed in the offseason.  The problem with this, of course, is that I've measured special teams using total points rather than, say, points per attempt (points per punt, points per kickoff, etc.).  Should we switch to measuring points per attempt?  Probably, we'll have to take care of that in the future.  Football Outsiders, where our motto is "methods subject to change at any time!"

## BUT WAIT, THERE'S MORE

Early on in 2003, I noticed that special teams were performing much better than the baseline numbers for 2002.  Were my baseline numbers way off?  Had special teams all improved dramatically in one year?  Or did weather, perhaps, play a major role in special teams performance that wasn't reflected in September numbers?  As the season went on, and special teams VOA ratings dropped for the league as a whole, it looked like the weather was definitely the culprit.  Just how much the weather affects special teams play will be an issue tackled in part two of our look at special teams.  We know from that kickoff list that stadiums matter for kickoff distance.  That article will also take a look at adjusting special teams ratings for opponent quality.  Alas, ye special teams fans, that article will also not appear until the offseason.  I've got bigger (playoff preview) fish to fry.

## THE FINAL RESULTS

This table shows the points over or under average for all 32 teams for each of the five aspects of special teams counted by Football Outsiders, along with the total.  In addition, you'll find the two "hidden indicators" of FG against and kickoffs against, along with a total that includes those numbers.  Ranks are given from #1 (worth most points above average) to #32 (worth fewest points above average).

 TEAM XP/FG R PUNT NET R PUNT RETURN NET R KICKOFF NET R KICKOFF RETURN ONLY R TOTAL R XP/FG VS. R KICKS VS. R TOTAL (incl. "Hidden") R KAN 2.3 14 4.4 13 18.4 1 3.0 11 16.1 2 44.2 1 -3.4 21 5.5 10 46.3 1 BAL 6.9 9 15.0 1 3.5 11 1.3 14 3.5 10 30.2 2 -12.8 32 15.2 1 32.6 2 OAK 9.6 5 11.2 3 16.9 5 -11.9 28 1.3 12 27.2 3 -7.3 28 6.1 8 26.0 6 PHI 4.6 12 -2.3 21 4.8 9 16.2 2 -0.5 13 22.8 4 6.2 5 1.7 15 30.7 3 CHI -4.0 23 5.5 11 17.3 4 -15.3 30 17.3 1 20.8 5 1.2 13 8.2 6 30.2 4 HOU 4.3 13 9.9 5 -3.3 22 10.6 5 -1.4 18 20.0 6 4.6 9 -6.6 27 18.0 8 PIT -10.1 29 8.0 8 18.4 2 0.4 16 -3.8 20 12.9 7 -4.4 24 8.4 5 16.9 10 CAR 7.4 8 -3.0 22 -4.0 24 2.8 12 5.6 7 8.8 8 10.2 2 7.1 7 26.0 5 NWE -10.0 28 -5.9 24 1.4 14 10.6 6 12.6 4 8.7 9 7.9 4 5.3 11 22.0 7 NOR -3.0 21 13.6 2 -1.6 17 5.0 9 -5.9 24 8.2 10 -0.5 15 -6.7 28 1.0 15 DET 13.0 3 -17.7 29 11.3 6 -1.0 20 2.0 11 7.7 11 2.7 12 3.9 13 14.3 12 GNB 6.9 10 -2.1 19 2.8 13 -12.0 29 9.0 5 4.6 12 -11.6 31 -0.2 21 -7.2 18 ATL -3.8 22 -0.7 17 8.0 7 21.3 1 -23.2 32 1.6 13 2.9 11 -14.8 32 -10.3 22 IND 19.1 1 4.2 14 -17.8 32 -9.5 26 5.5 9 1.5 14 -2.9 18 -6.3 26 -7.7 19 DEN 8.3 7 -19.6 32 1.2 15 16.1 3 -5.1 23 0.9 15 -7.6 29 -8.3 29 -15.0 24 NYJ 5.8 11 -19.4 30 7.5 8 -9.0 25 15.4 3 0.2 16 5.0 8 9.4 4 14.7 11 TEAM XP/FG R PUNT NET R PUNT RETURN NET R KICKOFF NET R KICKOFF RETURN ONLY R TOTAL R XP/FG VS. R KICKS VS. R TOTAL (incl. "Hidden") R CLE 1.2 15 3.6 15 -3.5 23 -0.6 19 -1.0 15 -0.3 17 -4.7 25 4.0 12 -1.1 16 SEA 0.2 17 5.8 10 3.3 12 -8.0 24 -2.4 19 -1.1 18 0.5 14 9.8 3 9.2 13 STL 17.2 2 -9.9 25 -6.5 26 10.9 4 -13.5 30 -1.8 19 -3.2 20 -10.8 31 -15.8 25 BUF -7.9 26 9.4 6 -2.6 20 -0.5 18 -0.8 14 -2.4 20 3.1 10 1.8 14 2.4 14 MIA -1.1 19 8.1 7 -8.1 27 -1.5 21 -1.1 17 -3.7 21 -4.0 22 0.7 18 -7.0 17 TEN 8.8 6 11.0 4 0.2 16 -11.4 27 -12.4 29 -3.8 22 -5.7 26 1.1 17 -8.4 21 DAL 0.2 18 2.3 16 -10.0 28 7.0 7 -4.3 22 -4.8 23 -4.4 23 1.3 16 -7.9 20 WAS 0.7 16 -0.7 18 -14.2 30 0.7 15 5.6 8 -8.0 24 14.7 1 10.3 2 17.0 9 NYG -6.0 25 -11.1 26 17.4 3 2.6 13 -15.2 31 -12.3 25 -1.5 16 -5.9 25 -19.7 27 TAM -11.0 30 7.8 9 -15.2 31 6.7 8 -9.4 27 -21.1 26 5.9 6 0.6 19 -14.7 23 CIN 9.7 4 -13.9 27 -5.8 25 0.2 17 -12.1 28 -21.8 27 -3.0 19 -1.8 24 -26.6 28 SDG -1.9 20 4.5 12 -2.0 19 -18.5 32 -4.0 21 -21.9 28 -8.9 30 -1.8 23 -32.6 30 SFO -17.3 31 -2.2 20 -3.0 21 -4.0 22 -1.0 16 -27.5 29 8.3 3 -0.4 22 -19.6 26 MIN -4.6 24 -19.6 31 -2.0 18 4.2 10 -8.6 26 -30.6 30 -2.5 17 -8.7 30 -41.8 31 ARI -9.4 27 -17.1 28 -10.8 29 -5.4 23 8.5 6 -34.1 31 5.0 7 0.4 20 -28.7 29 JAC -18.3 32 -5.5 23 4.5 10 -17.7 31 -8.1 25 -45.2 32 -7.2 27 5.8 9 -46.6 32