By John Ezekowitz
While other sports such as baseball and basketball have had statistical revolutions over the past decade or so, golf has been stuck in the stats dark ages. Watch a PGA Tour broadcast and you will see the same stats quoted over and over again: putts, greens in regulation, fairways hit, driving distance, total putting, sand saves, etc. Perhaps, if one is lucky, there will be a stat about a player’s average distance from the hole.
While these stats are important, just as say points are in basketball or batting average in baseball, they are often flawed indicators of success or efficiency (as those two aforementioned stats also happen to be). Now that PGA Tour Shot Tracker can give us information on literally every shot hit on the PGA Tour, its high time golf had its statistical revolution. Join me after the jump as I propose the first in a series of new stats for golf.
One oft-quoted stat that seems particularly in need of correction is driving accuracy, that is the percentage of times a player hits the fairway with his tee shot. As I showed in this earlier post, driving accuracy is not a significant predictor of scoring average on the PGA Tour. What this illustrates is that because PGA Tour players’ skills are all so finely tuned, the percentage differences in the amount of fairways hit by professionals do not predict scoring.
The players with the worst driving accuracy on the PGA Tour in 2009 hit the fairway about half of the time. Not coincidentally, the 10 worst in this category all drove the ball at least 298 yards, well above average for the Tour. The best players only hit the fairway about 72-74% of the time. That spread encompasses only about 2-3 fairways per round. The spread in driving accuracy for the PGA Tour in 2009 was not large enough to make it a meaningful stat.
Instead of the binary stat of driving accuracy, I propose that we track average distance from the center of the fairway. Since all new stats need catchy names, lets call this one Average Driving Deviation (ADD). This stat would allow for a much fuller picture of how tee shots affect scoring. Take, for instance, a player who hits only 55% of the fairways off the tee but averages 20 yards from the center of the fairway and a player who hits 70% of his fairways and averages 23 yards from the center of the fairway. Player 2 hits more fairways, but his misses are much further off line than Player 1. While Player 1 may find himself just in the rough more often, Player 2 probably finds himself in the trees further off the fairway or even in a hazard. On most golf courses, the further one is from the fairway, the harder it is to make a good score. On the PGA Tour, where saving a shot a round can make a player millions of dollars, those more wayward tee shots of Player 2 may cost him a shot or two a round.
Whereas driving accuracy’s binary nature can hide a player’s weakness–the variance of his drives from the center of the fairway– average driving deviation exposes how a player drives the ball, not just whether he hit the fairway or not. I believe that this stat would certainly be a significant predictor of scoring average, and would add quite a bit of new information about PGA Tour players’ skills.
Is anybody aware of any easily accessible large data sets for golf stats? Looking at shot tracker, it doesn’t appear you can export the data collected.
Measuring how far a ball is from the center of the fairway may be a better yardstick but I don’t know that its ideal. It doesn’t take into account that many times (almost every time in fact) Professional golfers aren’t aiming for the middle of the fairway. They are trying to be right or left because it allows a better angle to the green, takes a hazard out of play or fits a certain shot shape they prefer to hit.
Bit of a contradiction here:
Early in the article: “That spread encompasses only about 2-3 fairways per round. The spread in driving accuracy for the PGA Tour in 2009 was not large enough to make it a meaningful stat.”
Later: “On the PGA Tour, where saving a shot a round can make a player millions of dollars, those more wayward tee shots of Player 2 may cost him a shot or two a round.”
So are 2-3 fairways (and possibly shots) a round significant or not? I would say yes.
But beyond that, the author takes a good step: a pure average distance stat is not ideal in a game of accuracy. He also has a good idea to include deviation from the center of the fairway (accuracy).
However, we can take it farther: every hole in golf is measured against a predefined ideal: par. To make par, there is usually a well-known, identifiable, path on each hole, just like the perfect line in a NASCAR track: (drive 250+ yards to fairway, approach shot from ~150 yards to <20 ft from hole, putt~ 20 ft, putt< 6 ft).
The ideal stat would track a player's deviation from that perfect line. Golfers know that to make up for a bad shot, you don't have to make a good shot, but a great shot. Each player only usually has so many great shots in them a day. Calculating a player's slip off the line and how many times he gets back to it would tell us if he's in the zone, dodging bullets, or playing a historic round.