Reliability: The PGA Tour’s lost asset

LAS VEGAS, NEVADA - OCTOBER 18: Jason Kokrak of the United States looks on from the 12th tee during the final round of The CJ Cup @ Shadow Creek on October 18, 2020 in Las Vegas, Nevada. (Photo by Jeff Gross/Getty Images)
LAS VEGAS, NEVADA - OCTOBER 18: Jason Kokrak of the United States looks on from the 12th tee during the final round of The CJ Cup @ Shadow Creek on October 18, 2020 in Las Vegas, Nevada. (Photo by Jeff Gross/Getty Images) /
facebooktwitterreddit

From week to week, it’s hard to know who will perform consistently. Here’s why

How many of you had Jason Kokrak in the past weekend’s CJ Cup?

How about Martin Laird in the previous week’s Shriners? Or Sergio Garcia in the Sanderson Farms?

Unless you are related to Kokrak, Laird or Garcia, I’m guessing the answer is “not many.” Kokrak’s win,  after all, was his first in 233 career starts. Laird had gone winless in his most recent 168 starts before winning at TPC Summerlin. Garcia broke a 54-start winless streak at the Country Club of Jackson.

More from Golf News

That’s pretty close to normal on the PGA Tour, where week after week the fields are so competitive that almost literally anybody could win. Consider that the spread in 2020 seasonal scoring average between the best and worst players in the CJ Cup field was only about three strokes, with a standard deviation of a fraction more than half a stroke.

That is a statistical recipe for ‘anything goes.’

And that’s what makes picking winners on tour so perilous, a reality that ought to be of interest to anybody heavily into Draft Kings sorts of wagers.

The most reliable measure of forecasting any player’s future performance, obviously, is his previous performance. The problem is that Tour-wide there is only the most casual relationship between how a player does over an extended period of time and how he does on a week-to-week basis.

Since 2016, 75 players completed enough Tour rounds to qualify statistically in all five of those seasons. That group of 75 includes many of the game’s stars: Dustin Johnson, Rory McIlroy, Justin Thomas, Phil Mickelson, Justin Rose, Adam Scott, Jordan Spieth, Jason Day…you get the idea.

Yet the correlation between those players’ average scores and the week-to-week standard deviation of those scores is virtually random…less than 7 percent.

That lack of week-to-week consistency shows up constantly on Tour. Take our winning friend from this past weekend, Jason Kokrak. Since 2016, his stroke average based on 442 official rounds is 70.31. For the record, that is precisely the average of the 75 qualifying players over that time span.

But Kokrak doesn’t shoot that every week…not even close. In fact, only 105 of those 442 rounds ended in a score of either 70 or 71. That’s less than a quarter of his scores.

Measuring the standard deviation of Kokrak’s scores tells a different, fuller story. It is 3.25, slightly higher than the groups 3.22 average.

In civilian terms, while his average is 70.31, all we can reasonably count on from week to week is that Kokrak will shoot somewhere within 3.25 strokes of that average in either direction; that is, between 67.06 and 73.56.  You can think of those numbers as Kokrak’s and the group’s “reliability ratings.”

The problem is that statistically, that 3.25 stroke “reliability rating” creates a six and one-half stroke spread from high to low scoring possibilities. And there’s only about a two-thirds chance of Kokrak even landing inside that range. To achieve true reliability – say in the 97 percent range – we need to extend the data out a second standard deviation. For Kokrak, that’s between 63.81 at the low end and 76.81 at the high end.

For the record, in winning at Shadow Creek, Kokrak recorded rounds of 70, 66, 68 and 64. He had, by his standards, an extremely good week with two rounds (70, 68) in the range of one standard deviation below his usual performance, and the other two (66, 64) approaching the two standard deviation threshold.

But while Kokrak’s record demonstrates that such a week is possible, it also demonstrates that he could just as easily go the other way – 73, 75 for example – next week. How many times in the past five seasons has Kokrak turned in a round of 73 or worse? The answer is 108; that’s nearly one round in every four.

Next. Tony Finau Being Sued for Nearly $16 million. dark

The ideal player would combine a low stroke average with a narrow reliability rating. They would be the logical ones to bet on, especially if you were content to wager on a likely top 10 finisher as opposed to trying to pick an actual winner.

But are there any such players on Tour? Exploring that question requires an entirely separate article. Check back tomorrow to read about that.