A Rising Tide Can Sink All Ships
This week, the Tour takes us to Memphis, Tennessee, for the 2016 FedEx St. Jude Classic, played at TPC Southwind. The tournament has been at this course since 1989, so we have quite a bit of course data. Unfortunately, the data on this tournament is difficult to navigate.
At last week’s tournament, the Memorial, we had a very easy course. And although the scoring at Southwind isn’t dramatically down in comparison, our average Plus/Minus for golfers across all tiers is down significantly (per our free Trends tool). Take a look at the baselines:
At least $9,000: -6.50
$7,000 to $8,900: -5.17
No more than $6,900: -8.76
Compare those to last week’s numbers:
At least $9,000: +9.74
$7,000 to $8,900: +2.46
No more than $6,900: +3.84
Baseline matters a whole lot when evaluating DFS Golf stats. Scores in guaranteed prize pools will likely be lower this week than they were last week, which is fine because all golfers will likely have diminished scores.
A Smorgasbord of Variation
Now that we have our (not great) baselines, let’s add filters for our four important stats: Adjusted Round Score (Adj Rd Score), Greens in Regulation (GIR), Driving Distance (DD), and Driving Accuracy (DA). The following tables show how golfers for each salary tier have performed when above Tour-average for those stats. The first table is conditionally formatted together. The second is formatted by tier.
Normally, we see a distinct trend: Courses usually set up really well for bombers (emphasizing DD) or ball-strikers (emphasizing GIR and DA). This course does a little of both, depending on what salary tier you’re evaluating, which brings up an important question: For lower-tiered golfers, is ball-striking more of a proxy for talent than distance is?
Is There Even an Answer?
In the upper tier, we generally have guys who can do it all: They’re long and accurate. As you decrease in salary at DraftKings, you’re decreasing in overall talent, although golfers can still have elite marks in a couple of categories.
This article is designed to help find those golfers, whose talents are usually emphasized in the course. The problem is that there are some courses that aren’t black-and-white in terms of distance versus ball-striking. In those cases, we should emphasize talent.
However, if golfers are more-or-less properly priced and if as a result our best proxy for talent (Adj Rd Score) isn’t highly useful, then we might need to determine which of the other stats can serve best as a proxy for talent.
My guess here is that accuracy is more of a proxy for talent among low-tiered players than distance is. I admit that this is just a guess, albeit it’s an educated one.
The 2016 FedEx St. Jude Classic
Let’s walk through our stats now.
Adjusted Round Score: As usual, this is the best all-in-one statistic to measure the talent of a golfer. However, adding the Adj Rd Score filter for the top golfers doesn’t provide any value. They’re already all above Tour-average. However, we can probably find value by looking for cheap golfers with good Adj Rd Scores relative to their pricing. That is usually the best way to find value every week, and this week is no exception.
Greens in Regulation: This is where things get interesting and why I posed the question above. GIR doesn’t look helpful for the top golfers, but it’s incredibly important for the lower-tiered ones. More on this later in a second . . .
Driving Distance: Again, interesting. DD is important for the top golfers but not the lower ones. This is why I hypothesize that ball-striking is more indicative of overall talent than distance, as we see the biggest Plus/Minus bump in the lower-tiered golfers with great GIR marks. A lot of bad golfers can hit the ball a mile: Fewer can be as accurate as most PGA courses demand.
Driving Accuracy: To top off the weirdness, DA is highly negative for the high-priced golfers. Again, accuracy among the lower-priced golfers seems to be more valuable.
The 18th Hole
This week could be really tough for DFS players. If you’re using our Player Models, you might have to create two separate models in order to get proper ratings for each tier of golfer. Perhaps you should create one model that emphasizes distance and then see how it affects your high-priced golfers. And then create another model that emphasizes ball-striking and see how the cheaper guys are impacted.
Then combine those findings into (hopefully) winning lineups.
Good luck!
