Tuesday, August 6, 2013

On the problems of evaluating field goal kickers

 If you follow me on Twitter for some reason you may recall my objections to the general fan enthusiasm to cut Graham Gano or Billy Cundiff after every missed field goal or to praise Kai Forbath as a savior after every successful kick. You could be forgiven for thinking I was just being contrarian because I admit that's occasionally my thing. But I actually think the approach to evaluating field goal kickers - not only by fans and media but by NFL coaches as well - is deeply flawed in ways that will take more than 140 characters to discuss.


Let’s get right to the point: you can tell pretty much nothing about the quality of a field goal kicker from his field goal percentage. Aaron Schatz of Football Outsiders wrote back in 2006 that “There is effectively no correlation between a kicker’s field-goal percentage one season and his field-goal percentage the next.”

What are some possible reasons for this?

Maybe there simply is not a measurable difference in the field goal ability of the 32 best field goal kickers of the world.

If you or I tried to kick NFL field goals our success rate would most likely be zero, so clearly there is a drop-off in ability somewhere. But if the top, say, 100 kickers alive kick field goals with roughly the same ability, then of course we would fail to see any meaningful distinction among them at the NFL level.

Maybe there are differences in skill, but we just can’t measure them.

A full time starter in major league baseball may rack up 600-650 plate appearances in a season, and even with those sample sizes his batting average is subject to substantial variance based simply on statistical fluctuation. So when a placekicker attempts, say 35 field goals a year, a difference of a few percentage points in his success rate probably tells us absolutely nothing. Even over the course of multiple seasons or a career, there would be too much random fluctuation to make field goal percentage meaningful.

Maybe it is a Garbage In-Garbage Out situation.

Perhaps the stat itself is so meaningless that we shouldn’t be surprised it doesn’t yield useable real world results. Among the obvious weaknesses of field goal percentage:

1)   A 20 yard field goal counts the same as a 50 yard field goal, despite the fact that the difficulty levels are dramatically different.

2)   It doesn’t account for blocked kicks, bad snaps, poor conditions, and other factors beyond the kicker’s control.

3)   Perhaps the most pernicious: kickers who lack the leg strength for long kicks will not be asked to attempt them, so their success rate can be inflated because they are only trying kicks with a higher expected success rate in the first place. Likewise strong-legged kickers will be asked to try more 50+ field goals, and will obviously fail a substantial proportion of the time.

4)   A corollary of point 3: a kicker who had a poor success rate the previous year will likely be used less frequently as the coach may choose to punt or go for it more often when a field goal would be deemed difficult. Likewise a kicker who had been notably successful may be asked to make more difficult kicks, and his success rate would predictably suffer.  This could be mistaken for regression to the mean in the stats when it is actually due to external factors. (Note: I am only inferring that this effect would occur, but if someone with a stat background is feeling motivated it would make a pretty cool research piece if you could figure out a way to study it. Get on it, nerds!)

One way to control for some of these issues would be to divide up field goal attempts by range. To get some rough baselines for what sort of field goal success rates we should expect, I’ll steal some numbers from an old Pro Football Focus article:

“In 2010, kickers were perfect on kicks of less than 20 yards, made 95% of kicks between 20 and 30 yards, 88% between 30 and 40 yards, 74% between 40 and 50 yards, and 60% on kicks beyond 50.”

There’s probably some variance in those numbers from year to year, but what I’m doing here is just conversational rather than a statistical study so I will consider them close enough for government work (as the kids say).  We still haven’t helped ourselves all that much, though. Two big problems:

-         * If a full year’s worth of field goal attempts does not render a useful sample size for a kicker, then the single digit attempts in each range are even more useless.

-        * There is a “bucket” issue. Isolating field goals of between 30 and 40 yards (for example) is completely arbitrary and based on the numbers that happen to end in zero. I would guess there is a bigger difference in degree of difficulty between 31 and 39 yards than there is between 39 and 40, but that doesn’t get reflected here. If we go back through NFL history (or at least back to when the kicking style changed in the early/mid 80s) we could come up with an expected success rate for all 42 yard FGs, all 43 yard FGs, and so on (in fact, Advanced NFL Stats has done this). But again an individual kicker would never have enough attempts to let us know whether his performance at various ranges was significantly above or below average.

I think that by now I have established that you can tell next to nothing about a kicker’s field goal ability from his overall stats, and you certainly can’t judge him based on any one particular miss.


A field goal, by its very nature, is a probabilistic event. Anytime your kicker lines up for a field goal of between 40 and 50 yards, you can expect a miss about a quarter of the time.  If your team’s kicker misses and that kick happens to swing the outcome of a game, the fans, media and to a surprising extent the coaches are apt to question the kicker’s mental toughness and assign the miss purely to his alleged inability to perform in high pressure situations. This is absurd given that the guy had a 25% chance of failure just for walking out onto the field.

(Note: I do not doubt that the ability to perform in high pressure situations, and many other aspects of psychological makeup, have an impact on sports outcomes. But it is one of thousands of variable going into the result, and the habit of routinely holding it up as the only variable worth talking about is not only lazy analysis but also terribly unfair to the players who are being judged.  The mentally toughest athletes in the world will sometimes fail because the tasks that are asked of them are designed to be difficult – that’s pretty much the point of sports.)

And yet, here in DC we have seen a pattern play out that will be familiar in most other NFL towns as well:

-        * Shaun Suisham is kept around for several years due to his supposed reliability on field goals. When that reliability is called in to question by a missed chip shot field goal that leads to a 2009 loss against New Orleans, he is cut.

-       *  Enter Graham Gano, who despite a strong kickoff leg is shown the door after the 2011 season due to a mere 73.8% field goal rate as a Redskin.

-        * We then are treated to Billy Cundiff who continues the strong kickoff ability but is cut after going a mere 7 for 12 on field goals (58%!).

-        *  Finally we meet Kai Forbath, who saves the day with a 17 for 18 performance including an impressive 12 for 12 from 40 yards or deeper.

Consider the following facts and what they might say about the logic of the above series of personnel moves:

-        * Despite that last unfortunate miss against New Orleans, Shaun Suisham’s field goal percentage as a Redskin was 80.2%. The career success rate of Morten Andersen, considered one of the greatest field goal kickers of all time, was 79.7%.

-         * According to Pro Football Focus, Graham Gano and Billy Cundiff ranked first and ninth respectively for average kickoff distance in the 2012 season. Kai Forbath ranked 36th.

Given everything I have written above, there is little reason to think that Forbath will continue to make field goals at a much greater rate than his predecessors. He will almost certainly continue to be inferior on kickoffs, and those invisible yards matter (in fact the primary thesis of the Schatz article that I cited to begin this post was that kickers’ compensation should be based mostly on kickoff ability because that is the one skill in which they consistently differ from each other).

The last thing I want to do is to sell NFL coaching staffs short, so someone please correct me if I am inaccurately portraying their handling of kickers. But it sure seems like teams make decisions to cut their kicker and sign a new one based on the same results-based fallacies that lead fans and the media to conclude that a kicker is only as reliable as his last kick. With the exception of something obvious like a botched hold, whenever a head coach is asked about his kickers’ failures he will usually resort to something like “you just have to make that kick,” as if that means a damn thing. I get no indication that NFL kickers are regularly monitored and coached on their technique, which would actually give an indication of why they missed field goals and whether that result was bad luck or if the kicker is doing something that would indicate his struggles are likely to continue.


Consider this passage from the excellent book A Few Secondsof Panic, written by the sports writer Stefan Fatsis as he embedded himself in a Denver Broncos training camp and tried to convert his soccer skills to field goal kicking:

“… I learned the basics of kicking. Where to position my left, plant foot (about a foot to the left of the ball). How to take a backswing (left knee slightly bent, upper body straight, left arm extended to the side for balance, right heel reaching back to the right buttock, right foot pointed and locked like Baryshnikov’s). How to execute a proper downswing (snap the lower leg, keep the right foot locked and perpendicular to the body when striking the ball). How to finish the kick (straight at the goalposts).”

That’s a lot of technique; it sounds every bit as subtle as a baseball or golf swing. The tiniest flaw can throw everything off. So this leads me to what I have always thought is an obvious question:

Why don’t NFL teams have placekicking coaches?

In the book Fatsis hires a man named Paul Woodside who coaches instructional camps for teenage kickers. It is with Woodside – not a team employee – whom Fatsis consults on his technique and mental approach. If NFL coaching staffs, even the special teams coaches, are capable of detailed one-on-one instruction of their kickers, then I am unable to find public references to it.

Given that one or two missed field goals a year can mean the difference between a playoff appearance and a fired coaching staff, I am surprised that proper instruction and evaluation of kickers is not considered more of a priority.

Suppose a team had a guy on staff whose sole job was kicking coach. The advantages seem obvious. He could evaluate each missed kick for a true cause and instruct the kicker either on nuances of technique or mental approach. If no progress is seen the kicking coach can advise the head coach/front office that a change would actually be useful. If the correction does work he saves the team from reflexively cutting the guy and signing a stranger with no actual reason to believe the new guy is a better field goal kicker.  And by avoiding the unnecessary cuts, teams will not force themselves to settle for subpar kickoff performance based on a faulty or incomplete understanding of their current kicker’s field goal ability.

No comments:

Post a Comment