FanDuel PGA: Analyzing Trends in Optimal Lineups

Golf, the actual sport, is a really frustrating endeavor. When you work on your driving, your short game can suffer from a lack of practice time. You spend time on the practice green? Your irons may not be quite as sharp as they had been when you were at the range day after day.

Daily fantasy golf can be similar levels of frustrating, but we can certainly correct our thought process a lot easier than we can fix our slice or tendency to three-putt from 15 feet (can't be just me, right?).

One thing that can help is diving into historical optimal lineups to see which trends emerge that can help us think smarter about our lineup-building process.

So I ran optimal lineups from 109 PGA Tour events between 2018 and 2020 to see what stood out based on the perfect possible FanDuel lineup from each.

The Data

There's a lot to go over, so I'll start with the overalls, the general averages and such. I'll caution not to stop reading here because I get more granular. I'll add takeaways at the end if you don't want to read everything, but I, of course, think you should consider it.

Okay, so what this shows us is that, of the $60,000 salary cap on FanDuel, the average optimal lineup spends $57,251, which is actually pretty firmly below that threshold. Does that mean that leaving salary on the table is the right play? No, that's not really what that means. The more appropriate takeaway is that we don't need to stress too much about maximizing the salary when we build our lineups. We don't need to spend that extra $500 on a golfer we like less just to "use more of the cap." We should feel better about building lineups we want overall, within reason.

The average total lineup draft percentage sums up each golfer's draft percentage (e.g. if all six golfers drafted to a lineup were rostered at 10.0% across a contest, then the sum ownership of that lineup is 60.0%). That this number (68.6%) shows us is that rostering chalky players isn't always the right way to build a lineup, and that argument is supplemented by the average maximum draft percentage in these perfect lineups. The max draft percentage (i.e. the chalkiest, the most popular golfer in a perfect lineup) was on 28.3% of lineups. So we're not looking at the most popular plays (in the 30.0% to 40.0% range) making the perfect lineup at such a high rate that we must play them (I'll expand on that toward the end).

What's more indicative of the inherent randomness in golf is that the average winner's salary ($10,154) is very close to the average starting salary we have when building a lineup ($10,000) but that they are rostered at only around 15.1% on average. This would speak to a more balanced lineup build around some of the golfers with good-not-great win odds (e.g. neither favorites nor studs). This varies based on the field, of course. Sometimes we get fields with one or two superstars and little else.

Here's a breakdown of the averages per golfer (if we view golfers ranked by FanDuel points within the optimal, meaning the Golfer 1 (or G1) held the highest salary in the optimal, and the G6 represents the lowest salary in the optimal). The reason I'm ranking by salary and not FanDuel points is because FanDuel points are not easy to predict, and we're pretty much never rostering golfers at $9,500 with the anticipation that they will outscore golfers at $11,500.

Again, this is basically the average optimal lineup over 109 events if we bucket the golfers in the optimal lineups by their salary rank within the lineup. We do see that the FanDuel points correlate with salary, with the FanDuel points dropping as we go down in salary. The same is reflected in draft percentage numbers.

A possible overall takeaway here is that a double-stud lineup (roughly two golfers at or above $10,500) is very much viable, but that requires hitting on two golfers salaried near $8,000. That's just a general snapshot, but this doesn't show that the optimal trends toward three golfers in the $11,000 range and three in the $7,000 range; neither does it show that the average optimal is comprised of six golfers hovering around $10,000.

Events With a Cut

Does the data differ if we look just at events with a cut (most of which are "full field" events with nearly 150 golfers but not all), which make up the majority of PGA Tour events? Cuts and a full salary range can make a difference from a process standpoint.

Rather than compare the following data to what was already presented, it makes more sense to compare it to the no-cut data that we'll go over next. After all, 88 of the 109 lineups examined will be shown here, so we shouldn't expect this data to vary too much from what we looked at above.

So, this actually shows a drop in the overall salary and draft percentage from the above table. The reason for that? Full-field golf events are highly volatile (just try to bet them or simulate them -- it's really difficult to predict what happens over a four-round sample in golf when golfers are competing with a hundred-plus competitors). The numbers drop across the board, including the average winner salary ($9,966) in full-field events, which is actually slightly below the average starting salary ($10,000). We must be okay embracing the volatility and not always feel as though we need to jam in the betting favorites (who are almost always the highest-salaried golfers).

Let's look at the optimal lineup trend bucketed by salary in just full-field events.

It's about the same as the overall optimal trend: points and popularity descend with salary, and the build is virtually identical. Remember that 88 of 109 events I'm looking at qualify as cut events.

No-Cut Events

Let's now look at no-cut events, which range from roughly 30 to 75 golfers. We have 21 no-cut events to break down. Keep in mind that no-cut fields are overwhelmingly made up of elite golfers. (The average strength of field via the Official World Golf Rankings [OWGR] in these 21 events is 562.2; the mark in the 88 full-field events is 411.2. And OWGR field strength is calculated based on, in part, cumulative field size.)

There's kind of an inherent, ostensible idea that, in a no-cut event, we can punt with golfers at low salaries who are guaranteed four rounds of golf. This doesn't fully support that, but we'll learn more in just a second.

The average salary in a no-cut event's optimal lineup ($58,762) is higher than we saw in a full-field event ($56,891). Naturally (with fewer golfers to draw popularity), the cumulative draft percentage is higher in a no-cut event, and that trend is visible with the most popular golfer's draft percentage and the winner's draft percentage. Most notably, perhaps, is that the winner of no-cut events average around a full $1,000 higher than we saw a bit ago with full-field winner salary.

No-cut events aren't the right place to punt entirely; they do skew more top-heavy. I think it's vital to remember, though, that optimals rarely spend all $60,000, yet it does make sense to load up on two studs in no-cut events. The majority of points here are coming from the top end. You have to hit the right studs who separate in a no-cut event.

That said, that doesn't mean we should be locking in multiple golfers hovering near the $7,000 basement salary. Among 126 golfers in the no-cut optimals, only 11.1% had a salary of $7,900 or below. That compares to 17.2% in events with a cut.

Field Strength

It's never easy to capture the actual strength of a golf field, but we do have ways to estimate it, and we know that golf fields do vary in difficulty. Relying on the the OWGR field strength measures for these 109 events, I dug into see what stood out, if anything. However, because field strength does depend on things such as the number of top-200 golfers in the field, small fields and no-cut events will skew lower on the raw end. But we also know that no-cut events are basically all really tough, and even if we exclude them, we have 88 cut events to divvy up. So I did that.

Here is the overall data for events bucketed by field strength. Given that the 75th-percentile OWGR field strength rating my this sample was 560 and the 25th-percentile mark was around 211, I bucketed the events into difficult, moderate, and easy field strengths.

This shows heavy, heavy overlap between the 23 difficult events shown in the second column of the table and the 39 moderate events shown in the third column of the table.

We do, though, see a deviation when looking at the 26 easier fields, where the top golfer by salary comes in around $500 below the typical G1 in the other two types of fields.

This makes sense: the top golfers in a field such as the, for example, Safeway Open, may be the 80th-ranked golfer in the world but carry a salary akin to the top-10 when they are in action. There's a lot less certainty coming from the high-salaried golfers in weaker fields, generally.

Trends by Winner Salary

I've had a theory about how optimal lineups vary based on where the winner comes from in terms of salary for a while now. If a $12,000 golfer wins, wouldn't that probably lead to a differently structured optimal lineup compared to when an $8,000 golfer wins? I don't actually know, so let's find out. I'll look just at the 88 events with a cut so that we're not skewing the data too much.

Of the 88 lineups, 25 were won by a golfer with a salary of at least $11,000, and when that happens, the average total roster rate is 62.7% for a lineup. The winner's average draft rate is 21.6%. On average, 0.92 golfers with a salary of $8,000 or lower make the optimal lineup, and 2.44 golfers make it when salaried at at least $10,000.

We also have 27 fields -- so a similar split -- when golfers salaried at $9,000 or below win. The average winner salary is just $8,141, but surprisingly, only 1.6 golfers above $10,000 actually make the optimal lineup in this scenario. That speaks to the idea that if the stars (or at least golfers with the highest salaries in that field) aren't winning, they weren't good fantasy performers. Of note, 14 of these 27 events would've qualified as easy fields based on field strength, and only one was rated as a difficult event (the 2018 Memorial, won by Kyle Stanley).

How Chalky Picks Perform

A few years ago, I dug into how popular golfers fared (spoiler alert: not great) but have more data now and wanted to touch on it while I'm here.

There's an outright correlation between FanDuel Points scored and draft percentage of 0.31, which is fairly mild but far from zero. That suggests that drafters are able to identify some of the best performers fairly frequently. The correlation between FanDuel salary and FanDuel points, though, is 0.33, so it's basically as good (or better) long-term to roster golfers based on salary rather than the discourse surrounding him entering the week.

What that means is that when we're splitting hairs between Rory McIlroy or Justin Thomas in a given week, there's no evidence that we, as a collective, will be right in the long haul. That's a learning moment, though. We should embrace the volatility and embrace being wrong. If Rory is expected to be drafted by 40% of lineups but Thomas is expected to be on 25% of lineups, pivoting to Thomas is probably the right call.

That doesn't mean that we have to avoid every popular play all the time or to roster golfers with bad current form or course fit. But it's humbling to know that we're wrong, and it's advantageous to know that we'll be wrong often and leverage that into daily fantasy success.

Across the 109 lineups, the most popular golfer in the field has led the field in FanDuel points 9 times (8.3%). The most popular golfer in the field was a top-10 performer in 38 optimal lineups (34.7%). The most popular golfer was outside the top-25 in FanDuel points 38 times (34.7%), the same rate. So, we're as likely to see the chalkiest golfer to be top-10 as he he is to be outside the top-25, which makes him really not much of a factor in daily fantasy that given week.

Of note, the lowest salary that a most-popular golfer has held was $9,500; the average salary is $11,854. So we're really looking at good golfers, overall.

You know I'm going to divvy this up by field strength now to see what emerges. It's really different when we roster a $12,000 golfer in a major when we have a ton of other viable options, when we roster a $12,000 golfer who is one of three or four stars in a moderate field, and when we roster a $12,000 golfer who is 100th in the world but is at the top of the salary pool due to a weaker field.

A few things here: we see an inverse relationship with field strength and average draft percentage among the most popular golfer. This goes back to what I just said: in tough fields, we have alternatives that will take away some popularity.

However, we actually see the worst returns (based on average FanDuel point rank, top-10 FanDuel point rate, and worse-than-25th rate) when rostering the chalk in a tough field. That would apply to majors and should give you more confidence going against the grain in majors and tough fields.

The Takeaways

- The average optimal lineup spends $57,251, giving us wiggle room from the $60,000 cap.
- The average total draft percentage for golfers is only 68.6%. For example, that would suggest six golfers who are 11.4% rostered or a 30.0% golfer paired with five golfers at 7.7%.
- The average winner's salary across all optimals (and just in general) is $10,154, not that far from the average starting salary per golfer of $10,000.
- The average winner's draft percentage was 15.1%.
- Events with a cut (surprise) are more volatile than no-cut events, and we see the average optimal across 88 cut events spend $56,891.
- The average winner's salary in a cut event is only $9,966.
- The average optimal in a no-cut event is $58,762, higher than the cut optimal by $1,871.
- The total draft percentage in a no-cut event is 101.8%, due to the lack of options to be drafted.
- The average winner's salary in a no-cut event is $10,943, roughly $1,000 higher than in events with a cut.
- Even in no-cut events, just 11.1% of golfers in the optimals had a salary below $8,000. That compares to 17.2% in events with a cut.
- The average salary for optimal lineups coming from difficult fields is $58,097, compared to $54,588 in weak fields.
- FanDuel points correlate with draft percentage (0.31) a little less strongly than just with FanDuel salary (0.33), suggesting golfers at similar salaries perform similarly over the long-term and we should be okay avoiding significantly more popular plays in favor of golfers in nearby salary.
- The most popular golfer across all 109 events led the field in FanDuel points just 8.3% of the time.
- The most popular golfer in this sample was as likely to be in the top 10 in FanDuel points (34.7%) as he was to be outside the top 25 in FanDuel points.
- The average salary of the most popular golfer in the field is $11,854.
- The most popular golfer in difficult fields is drafted 33.5% of the time, on average, but finishes 39th in FanDuel points, on average. Such golfers are 60.9% likely to finish outside the top 25 in FanDuel points that week.
- That should give you confidence to avoid the chalk in a major or tough field.
- By contrast, the most popular golfer in weak fields is drafted 45.3% of the time and fares better, averaging a FanDuel point rank of 26.3 but is still 34.6% likely to be outside the top 25 in FanDuel points and just 26.9% likely to finish inside the top-10 (well shy of the 45.3% rate at which he's rostered).