About a year ago, I introduced the Vegas Bargain Rating (VBR) metric to identify discrepancies between DraftKings PGA salaries and odds to win a tournament. Such a metric is valuable because DraftKings weights player odds heavily in its pricing. Take a look at the correlation between salaries and odds for the U.S. Open:
The r-squared value of 0.87 is high and suggests that DraftKings prices players largely by their odds to win. That’s useful information because it’s not a perfect 1.0 correlation. There are outliers, and identifying those can help us find value in our quest to roster the winner of the tournament. And that’s useful because daily fantasy golf guaranteed prize pools (GPPs) are massive these days. If you want to take down a top-heavy GPP, you almost always have to roster the tournament’s winner. The VBR metric can help you find golfers who are cheap relative to their odds of winning.
To calculate VBR, I find a line of best fit (shown above), ‘predict’ what a player’s salary would be if there were perfect correlation, calculate the difference between predicted salary and real salary, and then reset everything to an easy-to-understand 0-to-100 scale.
And, of course, we can do the same exercise for golfers on FanDuel, where the r-squared value is lower than on DraftKings, but FanDuel VBR is still useful:
Without further ado, here are the U.S. Open VBRs for both DraftKings and FanDuel:
Site Differences
One quick note about the high-priced players with low VBRs: This exercise finds inefficiencies only within the posted odds. Jon Rahm has low VBRs on DraftKings and FanDuel of five and 13, but that doesn’t necessarily mean he’s a bad ‘bargain’; it means he’s bad according to his posted odds. You might disagree with his posted odds. I do. When interpreting this data, remember that it’s based only on salaries and odds.
If you look at the graphs above, you’ll notice a large difference between the r-squared values on DraftKings and FanDuel. It’s common for salaries and odds to win to be more highly correlated on DraftKings than FanDuel, but the different isn’t typically this large. While most weeks the difference is about 0.10 — DraftKings is usually around 0.80; FanDuel, 0.70 — this week it is 0.23. Relative to the past, DraftKings has a higher correlation, and FanDuel has a lower correlation. What are some potential reasons for this trend?
The pricing on DraftKings is fairly straightforward: Salaries come out early for major tournaments — they were released early last week — so it makes sense that they would rely more on initial odds. Additionally, there’s no history at this week’s course, Erin Hills. In fact, since opening in 2006, the course has hosted only two tournaments period: The U.S. Women’s Amateur Public Links Championship in 2008 and the U.S. Amateur Championship in 2011. Whereas most weeks there are years of course fit and history data to use when pricing players, this week there’s almost nothing at all. It makes sense that DraftKings would lean more on odds than ever this week.
So why is FanDuel so different? I’m not sure. It could be that FanDuel is pricing players more according to their talent with a metric like Long-Term Adjusted Round. Or perhaps FanDuel leaned more on recent play. I’ll dive into this further in tomorrow’s U.S. Open Breakdown, but for now note that the low r-squared value likely means that there are many ‘mispriced’ FanDuel golfers (if you trust the odds).
2017 VBRs
If you’ve been following this ongoing series for any amount of time, you won’t be surprised to see the studs with the highest VBRs on both sites. As usual, the sites cannot properly price these players given the salary caps. According to the lines of best bit, Dustin Johnson ‘should’ be priced at $15,377 on DraftKings and $16,724 on FanDuel. That isn’t realistic given the demands of roster construction, so he’s underpriced relative to his probability of winning this week’s tournament.
That means, as I’ve written before, that the most optimal way of rostering the winner of this week’s U.S. Open is to use a stars-and-scrubs approach. As mentioned above, it’s almost impossible to win big money in a GPP if you don’t roster the winner of the tournament. Look at the DraftKings Milly Maker:
In a field of 122,613 lineups, the chances of finishing first without rostering this week’s winner are statistically infinitesimal. That doesn’t mean that going stars-and-scrubs is the perfect way to build a lineup — there are other factors like course fit, ownership, etc. — but it does theoretically give you the best chance of rostering the winner. And that is perhaps the most important factor in PGA GPPs.
Of course, odds move throughout the week, and I expect to see some heavy line movement leading up to Thursday morning, especially if there seems to be a big weather angle with the wind. I will post a Part 2 of this piece with updated odds on Wednesday — be sure to check out our U.S. Open Dashboard throughout the week — but until then peruse the data in the table above and in our Tools, watch and listen to the videos and podcasts in our Premium Content Portal, and use our Lineup Builder to construct some GPP-winning lineups!
