So, we have to listen to you drone on about special teams now too?
Indeed you do.[1] At the end of the day, offensive and defensive performance are clearly more important factors in winning the game, but special teams is still a significant component of overall success — in fact, back in 2013 benevolent statistical overlord Bill C. indicated it accounts for about 14-15% of a game’s outcome.
1 | I mean, you’re already here, right?
And Charting the Tide, at least, wasn’t addressing it at all!
Back in December, we were all discussing how awesome J.K. Scott is[2], and I mentioned I would probably time punts this year. And I have. And I completely forgot I was going to do that when I wrote about this season’s new additions to the series. And I may not have remembered to do it until after the MTSU game. Whoops!
2 | Yes, is. Not as awesome, but still awesome.
At any rate, for the past two games I have been pulling the gross and net numbers on each punt and kickoff, as well as hangtimes, and I will continue to do that for the rest of the season. The latter is purely derived from game charting — you’re not getting that from play-by-play unless there’s a rather industrious statistician out there. The former two, of course, are tracked by the NCAA[3], and readily available at outfits like the wonderful CFBStats — as you’re undoubtedly aware, Scott had a higher gross punting average last season than any punter in the college or NFL, because Scott is the best punter alive.
3 | Well, returns are tracked, so you can back out net from that.
How I’m going to use that data will be a bit different, however. Instead of just giving you yardage, I’ll be using Bill’s Net Equivalent Points (NEP) model to assign a point value to each kick, which handily incorporates gross and net yardage into one number, thus doing a much better job of valuing a kick than either do on their own.
How so? Well, consider two punts from Scott in 2014, each for a gross of 60 yards. His 11th punt of the year came against Ole Miss, and was kicked from the Alabama 5. A great punt, but it was returned for nine yards to the Ole Miss 44. NEP values the field position gained by this kick at 3.7 points — below Scott’s average for the year, but above that of the Tide’s opponents. A solid kick.
A few games later against the Viles, Scott boomed his 25th punt of the year for 60 yards, but this time from the Alabama 24. After a return of negative four yards, the Viles found themselves on their own 12 yard line — not a great place to start a drive. That kick was Scott’s fourth-most valuable of the year at 5.94 points, over two points better than an identical-length kick from another part of the field.
I know what you’re saying — that’s great and all, but clearly the net yardage was different on those two kicks, which is where the value came from. Well, the kick against the Viles was good for a 64 yard net, 25% farther than the 51 yards on the kick against Ole Miss. The point value, however, was 60% greater, which suggests there’s something else going on here.
To wrap this up, consider two kicks of 41 net yards, and, coincidentally enough, 41 gross yards. The first, Scott’s 17th of the year against Arkansas, was kicked from the Alabama 44 to the Arkansas 15, with no return. The second, Scott’s 24th of the year and also against the Viles, was kicked from the Alabama 19 to the Vile 40, also with no return. The first was good for 4.05 points, the second for only 3.28. Not a huge difference, but a significant one nonetheless. Field position matters! This has long been captured with metrics like number of punts downed inside the 20, but that doesn't account for the value of punts to the opponent 21, 22, and so on. Rating kicks via NEP allows capturing the value of any kick, regardless of where it left and where it ended up.
Ok, ok, I believe you. What does any of this really mean, though?
So for each game this season, I’ll chart every kickoff and punt to produce an average hangtime, average gross points gained, and average net points gained, for both kickoffs and punts. The average gross will be a good valuation for the kicker, as that is completely under his control.[4] The net points gained will be a good valuation for the coverage unit when contrasted with the gross number, as it’s largely on them after the ball lands or is caught.
4 | Assuming he’s unmolested during the kicking process, that is.
So what are good numbers for these? Well, last season Scott averaged 4.54 gross points gained per kick, with a max of 6.98 and a minimum of 2.32. The Tide’s opponents averaged 3.95 gross points gained, with a max of 6.69 and a minimum of 1.22. In terms of net points, the Tide averaged 4.16 points, with a max of 6.34 points and a minimum of 1.60. The Tide’s opponents averaged 3.46 net points, with a max of 6.18 points and a minimum of -0.60 points — yup, you can go negative here given long returns.
Kickoffs are a bit different, since the vast majority are kicked from the 35 and make it to the endzone, and a large percentage result in touchbacks. A touchback from the 35 is good 6.62 gross points and 3.65 net points, which I think are good benchmarks for kicker and kick coverage performance respectively. Anything less than 6.62 average gross points indicates short kicks, and anything greater than 3.65 average net points indicates good kick coverage.
To that end, I will not include onside kick attempts or obvious squib kicks, since field position is not the goal of those kicks, and it would unfairly penalize teams by these metrics. As far as past performance, the Tide kickers[5] averaged 6.21 gross points, or about 62 yards a kick, and the coverage unit 3.68 net points, which when accounting for rounding errors, etc., puts the ball right at the 25 yard line. The Tide’s opponents averaged 5.99 gross points and 3.49 net points per kick last season.
5 | Scott kicked off several times when Adam Griffith was dealing with injuries last year.
So how’s it going this season?
Punts and Kickoffs Performance | |||
---|---|---|---|
Metric | ALABAMA | OPPONENTS | |
Punt Hangtime | 4.23 s | 3.52 s | |
Gross Points per Punt | 3.49 | 3.24 | |
Net Points per Punt | 3.26 | 1.80 | |
Kickoff Hangtime | 4.01 s | 4.15 s | |
Gross Points per Kickoff | 5.79 | 6.10 | |
Net Points per Kickoff | 3.97 | 4.43 |
I’ll note here that while these results are posted in a separate article this week, that is largely because of all the explanation required at the front of the piece. This table, or something similar to it, will be in the Defensive Charting the Tide moving forward. Just to clarify, I did go back and chart the kicks and punts from the Wisconsin game, so the above reflects results from both Wisconsin and MTSU.
Through two games, the Tide punter, punt coverage, and kickoff man are all lagging behind their performance in 2014, with the kick coverage unit a lone bright spot. Scott is not getting as much yardage out of his kicks, and that in conjunction with the loss of all-time punt gunner Landon Collins probably explains the performances there. Griffith is also not kicking off as far this year, but some combination of bad kick returners for the Tide’s opponents or good kick coverage puts the Tide a bit ahead of the pace in that regard.
The Tide’s gotten good punt returns off of bad punts through two games, which makes the opponent punting performance look pretty poor. In contrast, kick returns have been atrocious, and that’s reflected in stellar numbers for opponent kickers thusfar.
The hangtime numbers help explain some of this, as Scott is getting about 0.7s more out of his punts than the opponent, whereas Griffith is lagging behind by 0.14s. Hangtime allows the coverage unit to get down the field and prevent the setup of a return, so the more air you can get under the ball — especially on kickoffs — the better. Punts are trickier, as you want lots of hangtime, but not in a manner that sends the ball straight up in the air.
I didn’t time punts or kickoffs last year, and I’m not about to go back and try to do that now, but fortunately many people already spent a lot of time figuring out what a good hangtime is. Unfortunately I can’t find any of Paul "Dr. Z" Zimmerman’s great articles on the subject over at SI, but this website provides a benchmark of 4 seconds, which is what I recall from Dr. Z’s work back in the day. That applies to both kickoffs and punts, although with kickoffs you're going to top out in the 4.6-4.8 s range, usually — I believe the "record" is somewhere north of 5 seconds. Punts usually have a bit more hangtime, with 5 seconds being excellent and anything over that nigh-otherworldly.