SportGrid Radio SportsGrid
Golf PGA Premium Premium

PGA DFS Strategy: How to Play PGA DFS for Fun, Tilt, and (Hopefully) Profit

(AP Photo/Jason E. Miczek)
PGA DFS Strategy: How to Play PGA DFS for Fun, Tilt, and (Hopefully) Profit
Print Friendly, PDF & Email

This article will cover how to play PGA DFS for fun, tilt, and (hopefully) profit and break down concepts important to both seasoned and novice PGA DFS players. We will discuss important concepts for projections, current form, short-term form, course history, the dangers of strokes gained models, the efficiency of the PGA DFS market, and low vs high stakes concepts. We will cover leverage and Main Slate vs Showdown nuances. This won’t provide a foolproof path to profit, but will help shorten your learning curve, set you on the right path or help you avoid the wrong one.

How to Tease Out the Noise in “Picks” or “Projections”

Current vs Short Term Form

Like you, I have plenty of my own opinions about professional golfers. But realistically, we should not let them influence our PGA DFS play.

Watching Luke List miss a putt on the 18th hole to miss his 3rd straight missed cut in a row as your highest owned golfer will have a long-lasting impact on your mental well-being, but it is important to have a short memory in PGA DFS. But how exactly should you value that 3rd missed cut? Luckily, the kind folks at DataGolf have mathed out the projections we offer at DailyRoto including how to weight short-term form, current form, and course history.

You can read a lot about their logic here but the long and short of it is that they weight roughly 2/3rd of a projection on long term form, 1/3rd on short term form (last 3 months) and only a small sprinkle on course history (more on that below). The methodology is a lot more complex than that, weighing each progressively more recent round heavier, but at the end of the day, while our bias is heavy to what has happened over the last round or last event, the predictive value is stronger over a long time period.

Our default projections use DataGolf’s blend but there may be times where you want to make your own adjustments to our baseline projections. For example, if a golfer had a known injury in their long-term data, but has since recovered, I may want to weight their short-term form a bit heavier. For other golfers who have been off for more than a year (think Delaet or Kevin Chappell), you may want to ignore long-term form altogether.

Does Course History Matter?

DataGolf has probably done more work against course history truthers than anybody, but even they will acknowledge that course history does matter, with an asterisk.

They have a handy dashboard which illustrates that they have found course history to be more predictive at some courses than others. Course history needs to be reviewed with an understanding of that player’s baseline at the time of the event, and course history adjustments get stronger as a player has more rounds under their belt at that event. Our baseline has minor default adjustments for CH, but you may want to make individual player specific boosts to projections at events like The Masters (the strongest signal) while ignoring it altogether at places like Trinity Forest (among the weakest).

The strongest course history boosts could be as much as 0.3 strokes gained per round above a player’s baseline, while often times they will be non-existent. The other challenge with course history is that even if it was prevalent, would there actually be any edge?

Course history is almost always priced into betting markets, betting markets are almost always priced into DFS salaries, and course history almost always drives higher ownership from the PGA DFS community. That alone makes it something that is hard to derive an edge from.

In short, course history matters, but it is a lot more complex than how it will be discussed on your favorite podcast.

Strokes Gained Metrics (“Course Fit”)

If you listen to literally any PGA DFS podcast your brain is going to explode with strokes gained statistics. Even yours truly will be found guilty of spewing off anything from strokes gained T2G over the L50 rounds to SG APP over the L12 rounds as we add fodder to the airwaves. And the people that “nerd out” super hard will start spewing about SG splits on different grass types, putting surfaces and even attempt to crack Strokes Gained in different wind conditions with zero regards for sample size or field strength. I put “nerd out” in quotes because ironically it is usually the folks less analytically inclined who go down this path.

Thanks to the rising popularity of services like FantasyNational (from friend of DailyRoto Pat Mayo), it has never been easier to create your own custom report of strokes gained statistics over varying intervals of time. It is a fun way to research players, a good way to provide depth of commentary for podcasts, but IMO also one of the biggest traps a casual DFS player can fall into for trying to actually profit at PGA DFS.

The biggest issue is not that the data doesn’t have tremendous value (it does!), but rather that most people who use that information discuss the traits in a descriptive manner and not a predictive one.

They will tell you that strokes gained approach is important this week, or that the previous winners gained 45 percent of their strokes on approach, but often they will be citing the “course fit” based on their opinions without having backtested anything. They will never have reviewed a tournament by saying “Here is what the fields Strokes Gained metrics were leading into that event in a given year, and this is the predictive power each of them had on the final leaderboard.” They won’t have determined that while strokes gained putting represents 30 percent of the deviation in strokes gained at this course, that actually weighing it at 15 percent produces more accurate results. And they won’t help you make a predictive weighted stat model unless you do it yourself.

This creates false narratives that steam ownership onto individual golfers who may be at best marginally better values than those priced around them.

Since I know you are going to look at strokes gained metrics anyways, consider the following guidelines:

  • Review stats for golfers leading into the event rather than what they achieved at that event. This is the only data we have to access before predicting the next event so it’s important to leverage!
  • Strokes gained T2G is more predictive than strokes gained putting
  • Strokes gained putting (and around the green play) matters, quite a bit. Podcast sentiment often would lead you to dismiss it. However, it is much more volatile in the short-term. Consider using a longer time frame for putting in your models than Off the Tee or Approach play
  • Do not fall victim for short-term strokes gained splits stats… there are legitimately people who have cited strokes gained off the tee stats on courses with POA greens
  • Make small adjustments to projections on the basis of strokes gained metrics, not major ones

Analyze strokes gained data and by all means, you should join FNGC (after subscribing to DailyRoto of course) but be careful with how you implement it. Our projections factor in individual strokes gained performance with the same formula across all tournaments (with a heavier weight to tee-to-green performance than strokes gained putting) so if you find predictive course fit elements you can layer it on top of that.

The revised projection methodology for the 2020 PGA DFS season will include a course fit toggle that lets you transparently see the “fit” adjustments to our fantasy projections.

Why PGA DFS Can (and should) be Profitable

Simply put, PGA DFS has most of the dynamics that can make for a profitable DFS game. There is a high degree of casual money, limited public projection sources, our opponents are biased, and most importantly can’t beat salary-based expectations.

The biggest one for DailyRoto subscribers is that we are playing from a position of strength where we have better information than the market. Below is a view of the correlation between some key metrics (projections, actual GPP ownership at different buy-in levels, betting odds, and salary). This analysis was done for more than 3000 data points from full-field PGA events during the 2019 season. Far from an level sample size but enough to expect these to hold mostly true moving forward.

Actual ownership is correlated to DK performance but one of the weakest signals and is less correlated than simple salary-based expectation. To rephrase that: your opponents are no better at PGA DFS than salary-based expectations. This means that simple odds and salary based parameters would actually be good enough to compete and let others make mistakes.

Betting odds have a correlation similar to that of DFS ownership, with the obvious caveat that betting odds are missing the scoring elements (birdies and bogeys) that help improve projection accuracy.

As we scale up play, opponents seem to get worse. Ownership becomes congested at high buy-in levels in all DFS sports, but in PGA DFS it doesn’t seem to be identifying the best plays.

While correlation in PGA DFS is often limited except in extreme course fit or weather oriented scenarios, we should still have an edge. In fact, once you have controlled for salary, betting odds, and our fantasy projections, ownership actually has a negative coefficient in a linear regression model.

Our projections are materially better than the market and have a stronger correlation to results than salaries, betting odds, or ownership at any buy-in level. Does that mean every time Patrick Cantlay is projected as the best value that he’ll deliver? Nope. Does that mean the “top optimal” will be the nuts in cash games? Nope. But it does mean that DailyRoto subscribers are playing from a strong foundation in PGA DFS.


The upcoming season will be my 5th season playing PGA DFS, having started grinding PGA DFS since its inception and seeing the game evolve over the past four years. While my process and the industry have changed the one thing that hasn’t has been the underlying variance associated with the sport. While on a rolling 12-month basis I have been able to achieve roughly a 10 percent ROI there have been binks, peaks, troughs, and deep valleys that make you question the projections you use, your process and your sanity.

My RotoTracker graph looks roughly like someone having a heart attack while using a stair climber at the gym.

Each year I have been lucky enough to have one or two BINKs that carry my entire season. Without those, each year would flirt with breakeven. And the downside is worse. Despite a four-year track record of being profitable, there is enough variance in PGA DFS, and DFS GPPs in general, that having a losing 2020 PGA DFS season wouldn’t be entirely surprising. The market is always changing, usually getting smarter, and while I am a lifetime winning player, that isn’t an assurance that will extend for the remainder of my lifetime.

One last thing on the variance front, the PGATOUR changed their rules so that the cut is now to the Top 65 players and ties, opposed to the Top 70 and ties. This will lead to changes in PGA DFS, as fewer players will get all six golfers through to the weekend.

If you are signing up for PGA DFS, expect variance, choose a strategic process, and don’t let short-term results derail you from your path. Now onto the more PGA DFS strategy specific for large field GPPs.

This content is for members only.

Golf PGA Premium Premium