Fresh Take With MGFresh - NBA DFS Contests

Welcome to Fresh Take Tuesday, where we break down basic DFS concepts and offer a fresh take on going against the consensus strategy.  Most of the content here will relate to topics covered at times on the DG Courtroom Podcast.  However, the Judge can’t explain everything every day, he’s got to cover the daily slate, so we’ll cover strategy concepts here.

This week let’s dig into the different type of DFS contests and what that means for building lineups:

GPP (Guaranteed Prize Pool):

Large tournaments on FanDuel or DraftKings where multiple entries are allowed and winnings are typically heavily skewed to the top few scores.  A typical payout structure looks something like this:



To say GPP payouts are exponential is an understatement.  These types of tournaments pay out really well to the top 0.1% of lineups and ~80% finish out of the money.  Zooming in to smaller scale you can see the picture a little better.



By cutting out the best 99 lineups from this ~100,000 entry example, you can see that the majority of lineups lose money. The top ~20% typically at least double their entry with payouts increasing substantially if you get into the top 0.1% of lineups.  



When looking at the top 100 entries you see another hockey stick up. Meaning if you are in the top 0.01% of lineups in a large entry low fee tournament you are making thousands.  In this example from a $4 FanDuel tournament the best lineup is making almost 12% of the total pot or over 10,000x the entry fee.

Given this level of skew to the top lineups, typically DFS players will “swing for the fences” in these type of contests, playing the riskiest players to try and end up in the very small percentage of lineups that make massive sums.  A player taking this strategy has to be realistic that high variance players also increase your odds of being in the bottom 1%.  However, in GPP formats finishing in the bottom 1% pays the same as being just outside of the money cut off (~79th percentile) so this is a tradeoff many DFS players are willing to make.

Cash Contests (50/50 or Double Ups):

These contests from a payout perspective are very similar to each other but very different from GPPs.  

Technically in 50/50s half of the lineups lose 100% of the buy in and 50% return ~90% of the buy in.  The rest goes to the site as the rake for offering the contest. (GPPs have rakes too, it’s just not as easy to see in the payout structure.)

Whereas in Double Ups ~55% lose 100% of the buy in and 45% double their money.  (Those are round %’s so don’t @ me.)

In this structure the rake comes from more losers than winners.  Typically the total rake in 50/50s and Double Ups are similar.

In these types of contests there is no difference between having the best lineup and having a barely in the money one.  Therefore, players typically want to take less overall risk in their lineups and will select NBA players with “high floors.”  Typically these “cash plays” also have less upside or a less than ideal match up, but are consistently putting up a solid number of fantasy points for their given salary.  

Think of these players as having a range of points from 4x their salary to 5.2x with an expected score of 4.7x. These targeted ranges change from night to night based on the value offered on each slate, but for round numbers that’s a decent estimate.  This compares to “GPP” players who have lower floors and higher ceilings.  For example a “GPP play” may be very inconsistent and produce between 3x and 8x their salary.

Head to Head Contests:

Are exactly what they sound like…  Payouts are similar to a 50/50 contest, but you are only competing against one other lineup.  

One note of caution… If you blindly sign up for H2H contests you are likely playing against a professional.  Not that professional DFS players can’t be beat, but you’re likely missing out on the advantage of playing against less experienced players in other contests.  Strategy is typically similar to other types of cash games with less overall risk being taken. 

Single Entry Contests:

My personal favorite type of contest structure.  Payout weightings are similar to GPPs although the top prizes are less since there are fewer overall entries.  Unlike GPPs where professional players can enter 100’s of lineups to each contest, in single entry contests everyone plays on an even playing field with their single best lineup.  Here (similar to the Judge’s chambers contest) you can focus on creating your best lineup and you won’t have to worry as much about losing to someone’s 99th best idea that has some random ass bench players going off.

All other contests:

Typically some variant of the above types.  If you aren’t sure what a contest is, google them, or look at the payout structure on the “prizes” section of contest information tab to get an idea.

Fresh Takes:

The high risk/low risk thinking is fairly common in the DFS advice industry, but I think it can be taken too far by a lot of DFS players.  I believe your primary focus should always be on creating the best portfolio of players.  To do this I focus almost exclusively on expected points/$ cost.  This seems obvious, but I get the feeling a ton of DFS players are taking too much risk in GPP contests and not enough in cash games.

Building a portfolio of low risk and low expected return players in a cash game is a losing strategy. For example, building a lineup of players, each of which you expect to score 4.5x their salary with a floor of 4x is a losing bet if the average lineup with produce 4.8x their salary even though you played “low risk” players.  Conversely, you may have an option to play similarly priced high risk players with expected scores of 5x but a range of 3x-7x their salaries.  If one player is a bust (3x) and the other goes off (7x), as a package they produced a better total return than two low risk players everyone likes as “cash plays.” Try to think of the risk you want to take in terms of total portfolio or total lineup risk.

On the other side of the coin, in GPPs I see lineups with nine Thon Maker types who go off for huge scores once a year, but typically produce 3x their salary.  This seems like a losing strategy to me.  The odds of 9 very high variance players having their best night of the season, on the same night, on the night you selected them, is astronomically low.  Instead, if you have a few lower risk players in solid match ups, and take a few intelligent risks, you could have a giant score and differentiated enough lineup to win big money.  Again, the focus is first on expected point production per salary dollar with a secondary focus on adding a few risky plays which will determine if you end up in that coveted top 1%.

This is not to say you have to have the exact same lineup in cash and GPPs (although I personally think in a lot of cases you can play good lineups in both types of contests).  There are some situations which might make your lineups look very different.

For example, injury news.  It almost never makes sense to play a ‘game time decision’ player in cash.  If they don’t play, your chance of making money is nearly 0%.  In GPPs if the player has a good matchup and would be your top play if they go, then it may be a worthwhile risk.  Same if there is a backup which could play huge minutes if the starter sits.  In that scenario the backup is also a better GPP play than cash play, since their floor is reduced if the starter ends up playing. 

However, it is always a bad bet to play a low expected value player just because they will be “low owned.”  If Joel Embiid is questionable, and has a terrible match up and you decide to play him, consider your entry fee a donation to the eventual winner.

While the term “cash play” or “GPP play” is fine shorthand for low risk/high risk, do not take it too literally.  “Cash plays” can be in winning GPP lineups and GPP plays can make you money in cash. 


Now go get rich and check back next Tuesday for another Fresh Take. 

If you found this helpful please support DGCourtroom’s sponsors and the show.

You can follow Mike on Twitter @mgfresh