NFL Field Goal Probability Analysis

As I have posted, about the conference before, I will save you the Sloan Sports Conference rhetoric for the day except to mention there is some really, really, cool research being done by stats geeks to better understand sports. The topic today will be field goal kicking, and as most football fans know, the field goal kicker can be the game’s biggest hero or greatest goat depending on how well he kicks an oblong piece of leather through a 18 and a half foot opening. We could go on all day mentioning famous (or infamous games for field goal kickers, but to make things extra fun for us Auburn fans, ill mention one situation I feel is relevant to the paper I will be speaking about. This is the famous Florida game in the Swamp in 2007, where freshman Wes Byrum made the game-winning kick not once, but twice. On the road, in a hostile environment, game on the line, freshman kicker, timeout called just before he made the first kick, (forcing the re-kick) every situational factor fans think about as negatively affecting a kicker was going against Wes, yet he drilled the kick. This paper, presented at the Sloan Sports Conference a few days ago, details different factors influencing a field goal’s chance of being made. The link to the paper is found here: http://www.sloansportsconference.com/wp-content/uploads/2013/Going%20for%20Three%20Predicting%20the%20Likelihood%20of%20Field%20Goal%20Success%20with%20Logistic%20Regression.pdf

The paper intended to look at every possible factor which could influence whether or not a kick was made, including distance, weather conditions, situational conditions (score, time left in the game, etc. ) and psychological (playoffs, high-pressure kicks?). The data examined for the study was every single one of the 11,896 field goal attempts in the NFL regular season and playoffs from 2000-2011 (note: data found at http://armchairanalysis.com/nfl-play-by-play-data.php). From this data, the authors created a Binary Logistic Regression Model, which can be used to determine the probability a kick will be made while considering all of the factors which can influence the outcome. First, each factor was tested for significance, and it was found that distance was the greatest predictor, as expected, and most environmental factors significantly affected the outcome (except humidity). Interestingly enough, none of the situational/psychological factors were shown to have any significant impact on whether a kick was made or not. The researchers were then able to assign a coefficient to each factor, making it possible to enter each factor into a formula, along with the corresponding coefficients, to get an exact probability for any kick in any conditions.

Source: www.naplesnews.com

Using all this data together, the team set out to examine how good NFL kickers really were relative to standard measures of kicking success (kick made %). To do this, the team assigned a point value for every made kick and every missed kick. For example, a kick with a calculated .75 probability of being made would add 3*(1-.75)= +.75 points, while missing the same kick would be 3*(0-.75)=-2.25 points. By doing this for every NFL kicker, the team was able to account for kick difficulties in figuring out who the best kickers in the NFL really are. The reasoning is that degree of difficulty should be a factor in considering a kicker’s worth, not just raw field goal made percentage, because not all field goals are created equal. One of the names on the most underrated list shocked me, Mason Crosby. For those who don’t know, Mason has been the Green Bay Packers kicker for about five years now. As an avid Packers fan, I can personally attest that Mason was responsible for a few near heart attacks and temper tantrums over the past few years due to his mysterious inability to put the ball through the goal consistently. However, this advanced metric suggests that because he attempts a high number of difficult kicks, he is actually very underrated in his ability relative to absolute field goal make %. This is something which is difficult for me to accept, but they also rated the Packers’ home stadium as the 2^nd most difficult place to kick in the league, which could explain some of Mason’s misses.

 

Source: foxsportswisconsin.com (Mason misses another)

This is some very interesting research into the subject of field goal kicking, and with NFL games being much closer recently, every point is crucial. These probabilities should be used by a coach to make crucial kick/go for it on 4^th down decisions. He could have a chart made before the game for every possible scenario and know that if, say a kick had a 40% conversion chance but he felt his offense had a 45% chance of getting the first down, he should go for it. This type of forward thinking, although not always popular because coaches get skewered in the media for making anything other than the “safe” choice, should be a greater topic of conversation and research such as this will only further the discussion.