You Don't Need to Roster Volatile Players to Win a DFS Tournament
"Select volatile players in GPPs" has always seemed like a sound strategy.
But is it?
The conventional logic goes that while you should be maximizing your floor in cash games with safe plays like Jarvis Landry, you should be taking on more risk in GPPs, such as on boom-or-bust guys like Torrey Smith, due to a GPP's top-heavy payout structure. The problem is, many casual players take this strategy too far, frequently rostering players that can be classified as no better than a "dart throw".
How Do We Measure Volatility?
We can measure player consistency by calculating their coefficient of variation (CV) of fantasy points. CV is calculated by dividing the standard deviation into the mean, and the lower the resulting percentage, the more consistent, or less volatile, a player is, and vice-versa. For more on CV and positional consistency, check out my article on leveraging weekly player consistency in NFL DFS.
If the theory about selecting volatile players in GPPs is true, we should see players in winning GPP lineups with higher than average CVs -- the guys who have high the highest highs and lowest lows, so to speak.
How Volatile Were Players in Winning GPP Lineups?
It turns out that's not exactly the case. The tables below compare the average CV of players found in the grand prize winning lineups from the DraftKings Millionaire Maker and the FanDuel Sunday Million in 2015 to the average CV of all players at the position.1
Note: A negative figure in the +/- column indicates that the average player in a grand prize winning lineup was more consistent than the average player at his position.
|Pos||Avg CV in Winning LUs||Avg CV||+/-|
|Pos||Avg CV in Winning LUs||Avg CV||+/-|
For four of the five positions on DraftKings and five of the six positions on FanDuel, players in the grand prize winning tournament lineups were actually more consistent than the average player at their position. In other words, people aren't winning tournaments by choosing volatile, hard to predict dart throws. More often than not, they're actually choosing the boring, consistent players.
Let's now take a closer look, going site-by-site.
DraftKings Millionaire Maker
|Total - 11 Unique||16||47%|
The average DraftKings QB CV is 44%, and eight of the 16 QB selections were at 44% or below. Still, QB was the only position where the winning lineup CV was greater than the positional average, mostly due to Tony Romo's small sample CV. QB CV is relatively flat, so I'd focus more on finding a low-owned (but not necessarily inconsistent) selection here -- median ownership in the grand prize winning lineups was 6.9%, and only two slots had lower median ownership.
|Total - 27 Unique||40||62%|
Twenty-one of the 40 selections had a CV at or below the RB average CV of 63%. The important thing with RBs will always be volume. Even if an RB hasn't necessarily demonstrated consistency in the past, a large workload is a reliable indicator of fantasy production.
|Total - 32 Unique||56||55%|
WR is the position where you can really see the "select volatile options" theory turned on its head. The average WR CV on DraftKings is 60%, and 40-of-56 WRs, or 71%, came in below that mark. Furthermore, many of the most consistent WRs made multiple appearances in winning lineups. Of 56 WR selections (includes flex), only 32, or 57%, were unique. Eight WRs appeared in at least three lineups, and most of them are a who's who of consistency at WR: Antonio Brown and Julio Jones (5), and Brandon Marshall, Julian Edelman, Jarvis Landry, Odell Beckham, Larry Fitzgerald, and Allen Robinson (3).
Worried about ownership when rostering these consistent studs? Don't be. The median ownership for the highest owned WR in a grand prize winning lineup was 32%, and the average ownership for the second highest owned WR was 14%. There's another theory turned on its head: that you need a super contrarian lineup at every position. You need to mix in some low owned plays for sure, but you don't necessarily have to do it at all of your WR slots.
|Total - 11 Unique||16||65%|
Ten of the 16 TEs, or 63%, were at least as consistent as the average TE on DraftKings. In the case of TEs, the reason consistent players pop up in the grand prize winning lineups more often than not has to do with the fact that TE is a low-volume position, and most of the more consistent ones are also the highest-scoring ones, as they are more involved in the offensive game plan on a weekly basis.
FanDuel Sunday Million
|Total - 11 Unique||17||38%|
Eleven of 17 QBs matched or bested the QB average CV of 41%. The most volatile QBs in the grand prize winning lineups were only five percentage points higher than the average. As I mentioned earlier, QB consistency is relatively flat, I'd focus more on finding a contrarian selection here -- the median ownership of QBs in grand prize winning lineups was only 4.7% in 2015.
|Total - 22 Unique||34||57%|
RB on FanDuel had the second largest disparity between average CV in grand prize winning lineups and average positional CV. Seventy-four percent of RBs in grand prize wining lineups were more consistent than the average RB. The median ownership for the top owned RB in grand prize winning lineups was 22%, so you don't have to let potentially high ownership necessarily move you off a consistent play.
|Total - 27 Unique||51||57%|
Thirty-five of 51, or 69%, of WRs in grand prize winning lineups were at least as consistent as the average WR on FanDuel. Only 53% were unique. When you have roughly a one-in-a-million shot to win, it's incredible to see guys like Antonio Brown (6), DeAndre Hopkins (5), Julio Jones (4), and Odell Beckham all in at least one-in-four winning lineups. This is why I recommend paying up at two WR slots in GPPs. WR is a position where paying up gets you the most consistency relative to others at the position, as well as the most upside overall.
And again, if you're worried about ownership when rostering these studs, don't be. In the grand prize winning lineups, the highest owned WR had a median ownership of 26%, and the second highest owned WR had a median ownership of 13%. In other words, one chalk play and one moderately owned play at WR has been perfectly viable.
|Row Labels||Count||Average CV|
|Total - 11 Unique||17||60%|
At 10 percentage points, TE is the position with the biggest disparity between average grand prize winning CV and average positional CV. Only three times in 17 weeks was a TE in a grand prize winning lineup more volatile than the average TE on FanDuel.
The Bottom Line
Tournament grand prize winners aren't rostering a bunch of volatile players. Instead, they're rostering players that tend to be more consistent than the average player at his position. Furthermore, many of the consistent players appear in grand prize winning lineups over and over.
The DFS analyst code for "I have no idea how this (inconsistent) player is going to perform so roster him at your own risk" is "he's GPP-only for me". In reality, we should be spending less time analyzing who the perfect dart throw is, and more time analyzing which consistent players we are overlooking, especially at WR.
Now, of course, the methodology here isn't perfect, as we're measuring year-long consistency instead of consistency of a player up until the point at which he was in a winning lineup, but the takeaways remain relevant. Scan the lists of the players in winning lineups and you'll see a lot more players who exhibit average or above average consistency than you'll see volatile dart throws, especially for the players who have been in winning lineups multiple times. It's not that volatile players are completely absent from the grand prize winning lineups, but they do not appear as frequently as conventional wisdom might suggest. Deploy volatile players sparingly.
1. Included in the sample used to calculate the average positional CV were players who played in six or more games. Point per game cut-offs were as follows: quarterbacks who averaged 10.0 FanDuel points per game, running backs/wide receivers who averaged 8.0, and tight ends who averaged 6.0. To better capture true consistency, running backs who played as both starters and reserves had only their games as starters used when calculating CV.