The Making of StarsDraft’s PGA Product is proud to announce the release of our PGA product. As is always the case when we launch new sports there was a lot to consider. PGA is one of the least established of all the major season-long or daily fantasy sports. There’s really no set format or scoring system; the decisions change widely from platform to platform. On one hand this lack of focus can be frustrating — the fantasy golf enthusiast has to select from only a few relatively underdeveloped product offerings. For this provided a great opportunity to think critically about the unique problems facing PGA DFS and come up with solutions that result in the best product and most engaging experience in the industry.

Roster Setup = Draft 7 Golfers; Top 5 Scorers Count

The roster setup consists of 7 golfers, and only the top 5 scoring golfers contribute to your team’s final score. This provides some buffer against your golfers that miss the cut, and expands the strategic decision-making boundaries for our users during roster creation. As we’ve seen with the SUB position in NBA, this added strategic element is part of what makes roster creation at StarsDraft so entertaining.

Scoring System

Our scoring system will be entirely performance based. Points will be awarded as follows:

  • Double-Eagle: 15 points
  • Hole-In-One: 14 points
  • Eagle: 9 points
  • Birdie: 5 points
  • Par: 2 points
  • Bogey: -1 point
  • Double Bogey: -4 points
  • Triple Bogey: -7 points
  • Quadruple Bogey or Worse: -10 points

The only bonus points awarded will be in the event of a playoff at the end of scheduled tournament play. Holes played during the playoff won’t count towards a golfer’s final score, and the eventual playoff winner will receive a 5-point (one birdie) bonus.

We think this combination of added strategy, and more realistic performance-based scoring will make one of the most engaging DFS sites in the industry. Many of our users have expressed interest in how decisions regarding our games are made, and the analytics that underpin the process. So for the first time ever we are peeling back the curtain. What follows is an explanation of the data-driven reasoning behind our PGA product.

Making the Cut

PGA is unique in many ways from other fantasy sports, but perhaps one of the most interesting (and entertaining!) is ‘the cut.’ In most tournaments the cut is a mid-way point at which the bottom half of the field is eliminated from contention. Golfers that don’t make the cut are sent home Friday afternoon, never to play another hole in the tournament. Golfers that make the cut press on into the weekend playing two more rounds in an attempt to take home some hardware.

For fantasy purposes this makes golfers highly bi-polar. The worst golfer that made the cut is often much better than the best golfer that missed the cut (or they should be, if the scoring system makes sense). During roster creation the foremost concern becomes whether or not a golfer will make the cut, while the secondary consideration is whether or not they have a chance to do well on the weekend. One of the more frustrating experiences in all of DFS is to spend hours researching the perfect PGA roster only to have one or more of your golfers miss the cut. It generally automatically excludes you from cashing in most contests — and greatly reduces the entertainment value added by DFS for watching the weekend’s most pivotal rounds.

Given what has already done to address the frustrations of ‘Late Swap’ in the NBA, the PGA solution was obvious. We should allow users to draft a stable of golfers and play only a subset of them to increase the odds of having a complete team. We also don’t want to remove the skill of the game by allowing every team to have a complete roster. So what was the correct setup?

To figure this out I simulated 1,000,000 rosters for users of different skill levels at picking golfers that make the cut, under dozens of different roster setups ranging from a simple ‘pick 9 golfers, play all 9’ to a more complex ‘pick 6 golfers, play only the top 4’. A complete roster is one in which there are at least as many golfers that made the cut as are scored on a team. The plot below shows the results of this simulation.

Complete Rosters

The assumption for this plot is that our user’s skill at creating rosters with golfers that will make the cut is normally distributed around 60%. In other words, the assumption is that 60% of rostered golfers will make the cut. This might sound like a low figure, but keep in mind that users will be drafting rosters against a salary cap. So while there will be a McIlroy, Woods or Mickelson on each team, golfers that are very likely to make the cut, this will be offset by having to take at least a few golfers that have less of a chance.

After running the same simulation under several sets of assumptions about skill, we settled on a setup where users pick 7 golfers and play their top 5. In this construction — and given our 60% best guess about user’s ability to pick golfers that make the cut — 42% of rosters will have 5 or more golfers still in play going into the weekend. This gives a decent probability that a roster will be complete and have a chance to cash, while simultaneously preserving the much-deserved edge for users who pick a perfect 7 of 7 to make the cut. These users will be going into the weekend with what essentially amounts to a ‘free option’ where either of their 2 lowest scoring golfers could surge, skyrocketing their team to the top of the leaderboard.

With this innovation alone’s PGA product would represent a major improvement over the current DFS experience in terms of both excitement and strategic decision-making during roster creation. While the problem of ‘the cut’ is probably the first and most primary attribute that distinguishes PGA DFS from other daily sports, it certainly wasn’t the only consideration.

Daily Fantasy Scoring in PGA: What’s the Point?

Another important challenge in designing’s PGA product was the type of scoring system to use. Unlike our NFL, NBA or NHL products there were no sensible season-long scoring systems to aid our thinking. The most popular season-long PGA leagues are on Yahoo! Sports — a format where owners get an arbitrary 10 starts per golfer, and golfers are slotted into three different tiers for each event. Points are then awarded based on where each golfer finishes in each tournament. The best part about Yahoo!’s fantasy golf scoring setup is, well, that it’s some form of fantasy golf. Most golf fans and seasoned fantasy players agree that the ‘starts and tiers’ setup employed by Yahoo! is neither particularly engaging or innovative. Compared to the entertaining scoring rules in other major sports the ‘points for a win’ method of scoring fantasy golf adds very little excitement to the real-time PGA experience.

Translating this season-long fantasy golf into a daily format was impossible for’s purposes. First and foremost there are legal concerns; to qualify as a fantasy sport the outcome of each contest cannot depend on the outcome of a real world sports event. In the case of Yahoo!’s season-long golf, the only points awarded are for the real world outcomes of tournaments. For this reason, and because ‘points for a win’ scoring is just plain boring, DFS sites had to come up with another way to score fantasy golf. Though many variations exist throughout the DFS industry, we will focus on the largest DFS PGA site’s scoring for the purposes of this article. We will refer to this scoring as ‘bonus’ scoring for lack of a better term, and for reasons that will become obvious through this analysis.

How bonus scoring generally works is to award points for each hole, with progressively higher amounts of points awarded for lower scores and mild punishments for scores that are over par. Although there are plenty of variations, the most popular bonus scoring setup in DFS works like this:

  • Double-Eagle: 20 points
  • Eagle: 8 points
  • Birdie: 3 points
  • Par: 0.5 points
  • Bogey: -0.5points
  • Double-Bogey or Worse: -1 point

In addition to these scoring categories, you can be awarded bonus points if your golfer does one of the following:

  • Scores 3 birdies-or-better in a row
  • Gets a hole-in-one
  • They play every hole in a round at par or better
  • Finishes in the top 50 for the tournament

This scoring system is certainly a more exciting way to watch a golf tournament unfold than finishing-position-only based scoring. At we needed to consider this type of system first as the incumbent DFS scoring methodology and decide what, if anything, we should change. We strive to put a product that provides the purest fantasy experience possible — when the athletes you draft do well, so too should your fantasy team. As always, we turned to data to help uncover how such a system could be designed for the PGA.

Comparing DFS Scoring Systems for PGA

Something about the bonus scoring system jumped out at me immediately when I first began playing PGA DFS many moons ago. The relationships between scoring categories bared almost zero resemblance to the importance of various scores in actual golf. In bonus scoring, a birdie is worth 6 times as much as a par. Likewise, a birdie — or gaining one stroke with respect to par — is 6 times better than a bogey is bad.

Why is this notable? Imagine two scratch golfers both playing an even par round. Golfer A scores par on all 18 holes, giving him a total round stroke count of 72 and a bonus-based fantasy golf score of 18 * (0 .5) = 9 points (ignoring bonuses for the moment). Golfer B never plays a single hole to par — instead he oscillates between birdie and bogey every hole, ending with the same final round stroke count of 72, but with a bonus-based fantasy golf score of 9 * (-0.5) + 9 * 3 = 22.5. How is it possible that a 72 round of erratic bogeys and birdies is a 150% better fantasy performance than a 72 round consisting of only pars?

The insanity only grows if you stretch your imagination a bit and consider an even more schizophrenic golfer who oscillates between eagle and double-bogey for 18 excruciating holes to shoot 72. This maniac would be a fantasy all-star, putting up 9 * (-1) + 9 * 8 = 63 fantasy points in bonus scoring — a whopping 600% better than his boring par-only counterpart! And at the end of the round, the two golfers shot exactly the same score.

Any golf fan will tell you that the value of going birdie-bogey is identical to the value of par-par on any given holes; a stroke is a stroke. During many of the Majors, the Super Bowls of professional golf, the course conditions are such that even the best golfers in the world struggle to make par. The fundamental problem with bonus scoring is that it matters too much how a golfer gets their score, and matters too little what their score actually is. To resolve this conflict I’ve created’s scoring system — which I’ll refer to as performance scoring — as follows:

  • Double-Eagle: 15 points
  • Hole-In-One: 14 points
  • Eagle: 9 points
  • Birdie: 5 points
  • Par: 2 points
  • Bogey: -1 points
  • Double Bogey: -4 points
  • Triple Bogey: -7 points
  • Quadruple Bogey or Worse: -10 points

How does this system fare in our contrived examples? A golfer that shoots all pars for 18 holes is worth 18 * (2) = 36 fantasy points. A golfer that shoots bogey-birdie for 18 holes is worth, wait for it… 9 * (-1) + 9 * (5) = 36 fantasy points. They shot the same round with respect to par, so it stands to reason they received the same fantasy score.

The acute observer might have already noticed that this scoring system can allow for slightly different scores for golfers that happen to be more erratic — our schizophrenic friend’s round would be worth 9 * (-4) + 9 * (9) = 45 points: 25% better than his more pedestrian counterparts. Why is that? Shouldn’t all rounds of 72 result in the same score, if the scoring system were truly pure?

I think the answer is no. My reasoning is strictly mathematical. Good golf scores on a given hole are capped at 1 stroke — you can’t do better than a hole-in-one. Bad golf scores can be infinitely bad. As any true golfer will relate, golf is a merciless game. DFS is scored in real time, and the value of whatever is happening in that moment must be assessed immediately. An eagle actually does open up possible 18-hole scores that are unattainable to someone who only scores birdie or par — the closer you get to the perfect score of 1 stroke on each hole, the better your eventual final 18-hole score could possibly be. If you’ve ever shot a low round on the front 9 of a golf course you’ll know exactly what I’m talking about. The fact that you went low on the front 9 offers up the daunting potential for a phenomenal final score, which is exactly why you shoot 12-over on the back. So there are marginal advantages, in the moment, to having shot an eagle over having shot a par, even if the eagle ends up being counteracted by over-par scores later on in the round.

So far this may have all seemed like gibberish. “Who cares!” you might be tempted to shout. “If you pick the best golfers you’ll still win all the money!” Some things can only be expressed with an image, so I produced the plots below showing the distribution of fantasy scores for real PGA rounds at various stroke thresholds over the last 4 years:


In the mess of distributions that is bonus scoring you’ll find a 10-over par round that was worth more fantasy points than many even par rounds. You’ll find another 6 under round that was worth less than the average 3-under round. The overlap of the distributions is a sign of just how impure and random the bonus system is. Contrast this to the distributions created when we apply performance scoring outlined above — each distribution is cleanly separated out by the total round score, with slight deviations upward for golfers going lower than birdie on multiple holes. Rounds of 3 under par are clearly and consistently worth more than rounds of even par, and rounds of 3-over par are clearly and consistently worth less. The performance plot makes sense; the bonus plot does not.

But a pretty picture is not a good reason in and of itself to change the way scores a sport, even if it is based on solid data. I wanted to know what impact this apparent madness has on actual fantasy outcomes: what percentage of the time would a 3-over round outscore an even-par round? Did a 3-over round ever outscore a 3-under round? How about a 6-under round? And what would these numbers look like in the new performance-based scoring? I was taken back by what I found:

Daily Fantasy Golf Plus Three Grid

In bonus scoring, 64.2% of 3-over rounds outscored the lowest scoring even-par round recorded. 18.4% of 3-over rounds outscored the average even-par round. Amazingly, 5.6% of 3-over rounds outscored the average 3-under round! Imagine having selected a golfer that posts a 3-under round and being outscored by some 3-over hacker — oh, the DFS agony! And it didn’t stop there:

Daily Fantasy Golf Plus Six Grid

In my mind this all seemed to make sense and provided a reasonable argument in favor of performance-based scoring. It was still possible that I was being misled by the data, however. Maybe what I was seeing was the result of taking rounds from different courses or tournaments too far out of context. Maybe comparing a 3-under round from the Masters with a 3-under round from the John Deere Classic is somehow distorting the data and misrepresenting the characteristics of these two scoring systems. I wanted to understand how the systems compared at the level of the tournament.

To visualize this I looked at hundreds of tournaments (and selected 16 of them at random for this article) and plotted each golfer’s finishing place in the tournament against their fantasy points scored in the two systems. These graphs only include golfers that made the cut at these events and points scored based on hole-by-hole points, including bonuses for 3-birdie streaks, par or better rounds, and hole-in-one bonuses. I’ve removed the axes of each plot for simplicity. The point of these graphics is to visualize what we think should be a correlation between finishing place and fantasy points earned in each event:

Daily Fantasy Golf Tournament Grid

The above plots are probably confusing — many of them just look like an uncorrelated cloud of points. That’s because they are uncorrelated clouds of points! For bonus scoring, even within a given tournament, the correlation between fantasy points and finishing position is typically very low. This is especially true in events on easier courses that yield many low scores. Compare this to the same plots in the performance scoring system:

Daily Fantasy Golf Perf Tournament Grid

The clear S-shaped curve for almost every event corresponds to the normal distribution of golf scores in these tournaments. The winner and the top-5 or top-10 golfers typically outperform a large group of middling golfers, who themselves outperform a handful of golfers that make the cut only to experience golferia, imploding on the weekend and putting up terrible fantasy numbers. Again, performance-based scoring makes sense, bonus scoring does not.

Bonus Bonanza

Okay, so I think we have solid evidence that bonus scoring doesn’t consistently award the most fantasy points to those who play the best golf. How then does it persist as the scoring system for most of the DFS PGA world? The bonus system has a major Band-Aid: finishing place bonuses. These range from 30 points (1st place) to 1 point (50th place). The intent of this final-place bonus is to ensure that despite the madness of bonus hole-by-hole scoring, the golfer who wins the golf tournament is also the golfer with the best fantasy score at the end of the tournament. This sounds great in theory — if the hole-by-hole scoring diverges from reality, adding a bonus to higher-finishing golfers can make up the difference. But when you look closer, there are several problems with this approach that ultimately detract from the experience of playing inside a bonus scoring system.

The first is that, as with all fantasy bonusing systems, the point values are arbitrary. Why is winning worth 30 points? Why not 50 or 10? Is first place really worth as much as 60 pars or as little as 3.75 eagles, or 1.5 double eagles? Second place is awarded 20 points, a 10-point drop off from first place. In Tiger Woods’ heyday, maybe this made sense — when he won, he often crushed the field — but even then many tournaments came down to a playoff, and many playoffs were decided by a putt that veered a quarter of an inch one way or the other. In other words, even though not all second places are created equal, in bonus scoring a second place is always worth 10 points less than a win … and 10 points is worth more than an entire even-par round!

One of the most resounding points of feedback I received when talking with PGA DFS enthusiasts about our eventual PGA offering was a desire to decrease the emphasis on place-based bonusing in’s scoring rules. My intuition was to get rid of bonuses entirely; if you think out your scoring system well enough you shouldn’t need an end-of-tournament adjustment to make sure fantasy points align appropriately with performance. To a large degree I think performance scoring as outlined above does that, but there’s one very important catch. In golf there are often ties at the top of the leaderboard at the end of an event. These ties result in playoffs to determine the eventual winner. Winning a playoff requires skill, and that skill should be reflected in a golfer’s fantasy score at the end of a contest.

Unfortunately the holes played during the playoff can’t be included in fantasy points — playoffs vary in style and duration. For example, the U.S. Open goes so far as to require an additional 18-hole round of golf to determine the winner. Other tournaments have sudden-death style playoffs. In order to adequately account for playoff performance, we’ve added a 5-point (one birdie) bonus to any player that wins by way of a playoff (a bonus that isn’t granted for those who win a tournament outright).

So that’s it. We’ve gone from an ill-shaped concept of what fantasy golf should be like to the final specifications of’s PGA product. The rosters will consist of 7 golfers, and scores will be calculated from the top 5 to minimize the frustration of having incomplete rosters going into the weekend. The scoring will be performance-based, offering the most realistic and pure real-time experience of a PGA event — with the only bonus being for a playoff win. Creating a new DFS offering from scratch is a mix of art, science, and common sense, and wanted to pull back the curtain on our thought process for PGA to help our users get a glimpse of all that’s involved. Good luck in your contests, PGA or otherwise, and please let us know your thoughts by emailing:, or follow us on twitter @StarsDFS.