Strokes Gained-Putting revolutionizing golf statisticsDecember 16, 2011
PGA TOUR staff
First, consider the possibilities ...
ShotLink, powered by CDW, the PGA TOUR's official technology partner, has been around for 10 years, collecting and cranking out real-time data on every shot of every player. Which, using some not-so-simple math, amounts to: an average of 450 rounds per tournament, which amounts to approximately 32,000 shots; times 40 ShotLink events equals 1.3 million shots tracked per year; times 10 years equals 12.3 million shots collected. Then, with six "attributes" collected per shot (lie, location, distance, etc.), it amounts to 77.8 million attributes that have been recorded through 2011 to compute everything ShotLink is able to deliver
Then, to offer that much raw statistical data to those academic minds, it was bound to happen, coming up with something so bold and new. But, then, that was the idea when the PGA TOUR quietly began offering ShotLink data to academic institutions for research beginning in 2005.
In this particular instance, it was the brainpower from Columbia Business School and Massachusetts Institute of Technology that devoured an endless stream of putting numbers tracked by ShotLink and, unconfined by the traditional view of golf statistics and what they're supposed to tell us, analyzed it in a most unconventional way.
Putting stats to that point had been rather unsatisfying in terms of becoming a meaningful gauge to identify the best putters on the PGA TOUR. While ShotLink had been providing a wealth of putting data since its official introduction to determine proficiency from various distances, the primary overarching putting statistic continued to be Putts Per Round. Pretty standard for golf, but it simply measures the average number of putts a player takes over 18 holes and can be skewed by chipping close to the hole after missing a green.
Professor Mark Broadie of Columbia originally came up with an alternative concept of "strokes gained" through putting. Then, in 2009, Graves and a team from MIT used Broadie's "strokes gained" concept to rank putters on the PGA TOUR. Finally, in 2010, the PGA TOUR refined the formula and implemented Strokes Gained-Putting as the PGA TOUR's new primary statistical category in May 2011. To rave reviews, it might be said.
As a basic overview ... Strokes Gained-Putting is computed by calculating the average number of putts a PGA TOUR player is expected to take from every distance, based on ShotLink data from the previous season. The actual number of putts taken by a player is subtracted from this average value to determine strokes gained or lost. For example, the average number of putts used to hole out from 7 feet 10 inches is 1.5. If a player one-putts from this distance, he gains 0.5 strokes. If he two-putts, he loses 0.5 strokes. If he three-putts, he loses 1.5 strokes. A player's strokes gained or lost are then compared to the field. For example, if a player gained a total of three strokes over the course of a round and the field gained an average of one stroke, the player's "Strokes Gained Against the Field" would be two.
All that really needs to be known is that the players totally buy into it as an accurate, meaningful gauge ... that, and the fact that it demonstrates the possibilities when putting that much data in the hands of very smart people. It's pure fodder for creative, analytical minds.
This program that established a formal process for professors and students to enter into a license agreement to gain access to ShotLink data today is called ShotLink Intelligence Powered by CDW. CDW has been the PGA TOUR's technology partner since 2008. And while ShotLink was developed before the official marketing partnership with CDW was established, CDW has played a significant role in upgrading, streamlining and enhancing the mechanisms of ShotLink.
ShotLink Intelligence has grown substantially since those early days of the program and to date has made the data available to faculty and other participants from more than 65 institutions, and that number continues to grow. A sampling of institutions represented includes: Carnegie Mellon University, Columbia University, Georgia Tech, Harvard University, Johns Hopkins University, Massachusetts Institute of Technology, Stanford University, University of Chicago, University of Michigan, University of North Carolina, University of Pennsylvania and Yale University.
What's interesting, though, is that analysis hasn't been confined to golf. Several research papers have applied golf data to business. An example of this was authored in April 2010 by Jennifer Brown of Northwestern University's Department of Management & Strategy and entitled Quitters Never Win: The (Adverse) Incentive Effects of Competing with Superstars.
For the Abstract, she writes:
Managers use internal competition to motivate worker effort, yet economic theory suggests that the benefits of competition may depend critically on workers' relative abilities -- large differences in skill may reduce competitors' efforts. This paper uses panel data from professional golfers and finds that the presence of a superstar in a rank-order tournament is associated with lower competitor performance. On average, PGA (TOUR) golfers' first-round scores are approximately 0.2 strokes higher when Tiger Woods participates, relative to when Woods is absent. The overall superstar effect for tournaments is approximately 0.8 strokes. The adverse superstar effect increases when Woods is playing well and disappears during Woods's weaker periods. There is no evidence that reduced performance is due to "riskier" play.
She makes her case over the next 28 pages.
Again, it's just one non-golf example. Of course, there are many papers that take a refreshing new look at what ShotLink data can tell us about the PGA TOUR and the players. And there will be many more as the program continues to grow.
After all, 77.8 million attributes are just the tip of the iceberg. Who knows what the next 77.8 million will tell us?