There are countless resources out in the DFS world on lineup construction, GPP strategy, top plays, and ways to help you figure out how to construct a profitable lineup. But when it comes to what contests to put those lineups in, the research is pretty bare. Yes, we certainly have a good idea which are likely the highest ROI contests to enter, but are those contests going to make you the most profit? When we decided to take on this project of solving contest allocation, we thought the answer to this was quite simple, but it turns out it’s not. And most of the advice you’ve heard? It’s actually wrong.
We’re going to just get right to the point. Here is what you should be thinking about when entering contests.
You SHOULD NOT maximize for ROI when selecting contests.
You SHOULD be maximizing for profit.
In order to maximize your profit, you need to be minimizing your variance as much as possible. By doing this, we can risk a greater part of our bankroll per day without risking going broke, and therefore make a higher profit over the course of a season.
A mixed contest allocation strategy (we will expand upon this more later) is the best contest allocation strategy for making money.
Lineup diversity in underrated and underutilized. Using a wide variety of lineups will lower your variance and therefore make you more money in the long run.
The flagship contest (in MLB DFS, this is the relay throw on DraftKings and the squeeze on FanDuel) is a great place to gamble but a terrible place to build your bankroll.
In this blog, we will go over our process we used to come to these conclusions and the evidence we found to support our findings.
The Research Process
When we first took on this project, we thought we could simply look through the DraftKings contest histories we had in our database (FanDuel doesn’t support easy access to these results) and come to some good conclusions about contest allocation. But it was hard to find a way to take these results to actually mirror the grind of a DFS player. So we decided to leverage SaberSim’s own MLB simulator to create what we dubbed a full contest simulation.
Without getting too technical, we essentially created portfolios of lineups for different MLB slates, based on actual DFS players lineups, and then used 3,000 of our own simulated results for each slate to project out how these lineup portfolios would do over many different stretches of time, kind of like DFS alternate universes. This allowed us to truly see the actual experience of DFS players, and how that their bankrolls rise and fall over time.
When we organized this data, we discovered some really useful things for contest allocation.
The Profitability of Non-Experienced Contests
With our full contest simulator, we were able to specifically quantify the difference between normal contests and non-experienced or veteran-excluded contests. We took contest pairs, for example the single entry $12 base hit and the single entry $1 daily dollar or the $4 20-max Four-Seamer and the similarly structured .25 quarter jukebox, and we took all the lineups from one contest pair and put them in the other contest, then simulated out that contest 3,000 times. We did this over several different slates, and we found that in general, the lower buy-in, non-experienced contest has a 2-4% ROI difference from the higher buy-in contest.
The most clear illustration of this was with the $4 Four-Seamer, a 20-max contest that DraftKings runs daily. On many days, this contest fills and DraftKings creates another Four-Seamer that is non-experienced because of prize pool rules. When we took the experienced Four-Seamer lineups and put them into the non-experienced Four-Seamer, they did almost 4% better in ROI over our simulated outcomes.
This information was probably something you already suspected. When the best players in DFS have access to a contest, they will probably enter it, and those players are going to hurt your ROI the most. But we were excited to be able to back this one up with hard evidence.
Single Entry and 3-Max Contests: The Highest ROI and the Highest Variance
It’s common wisdom among DFS players that single entry and 3-max contests are some of the highest ROI contests in DFS, and our research (unsurprisingly) confirmed this. But as we looked deeper into contest allocation strategies, we discovered something shocking: exclusively playing these contests actually makes you less money in the long run.
The reason for this is variance. Using our contest simulations, we experimented with the results of different contest allocation strategies over a simulated DFS baseball season. We tried single entry and 3-max exclusively (we call these “elevator” contests), multi-entry exclusively (we call these “diversifier” contests), and different forms of mixed allocation strategies. As it turns out, an elevator-exclusive (we’ll abbreviate this to “SE”) contest allocation strategy has a lot of variance, even if you are using several different lineups per contest.
Let’s take a SE strategy wagering $100 per slate. Over our contest simulations, we found that with a $3,400 bankroll, we will bust our entire bankroll around 25% of the time. For a mixed entry strategy, entering mostly diversifier contests, we could have a $1,200 bankroll and bust our bankroll at the same rate. In fact, using a mixed entry strategy, allocating 70% of our entries per night to diversifiers and the rest to elevators, we can actually risk $250 per night with only a $3,200 bankroll and bust at the same rate.
The profit results of being able to risk more per night with the same bankroll is staggering. Our SE strategy with our $3,400 bankroll has an end of season expectation of around $1,800. But the mixed strategy with a $3,200 bankroll? Our end of season expectation is $6,240. We make almost 3.5x the amount of money by simply using a better contest allocation strategy.
Keep in mind that our SE strategies in general utilized 10+ lineups per slate. Using only a few lineups, or just one, is going to be a much more high variance strategy and you are going to bust your bankroll at a much higher rate than our numbers suggest. So if you’ve already busted your bankroll this MLB season, it may not be a skill issue, it may be a contest allocation issue.
The Variance of the Flagship Contest
Both DraftKings and FanDuel run what we call a flagship contest every day for a given sport. In NFL, this would be the Millionaire Maker or Sunday Millions. In MLB, these are usually the Relay Throw on DraftKings and the Squeeze on FanDuel.
These contests have two things that make them very attractive:
They offer a large amount of money for 1st place. Sometimes over $50,000 or $100,000.
They have 10,000+ entries every day.
But while these may be a fun gamble, our contest simulations showed these are horrible for building your bankroll. We found that for 5-20% ROI players, maxing these contests out with 150 entries, profited approximately once every 6 days, and needed almost a 100,000 bankroll in order to avoid going bust.
This is because these contests skew their prize pools heavily towards first place, and this makes the variance absolutely insane. Because of this, we actually excluded using these contests in our allocation strategies.
So how should we use all this information? Here are our actionable recommendations:
Use 2.5-5% of your bankroll per day.
Enter both diversifier and elevator contests.
Diversifiers are 20+ entry max contests other than the flagship contest.
Elevators are single entry and 3-max contests.
Use 50-75% of your daily wager on diversifier contests, always prioritizing the lowest entry fee and veteran-excluded contests first.
Use the rest of your daily wager on elevator contests. Again, always prioritizing the lowest entry fee and veteran-excluded contests first, and working your way up.
It’s easy to imagine why the common wisdom for building a bankroll in DFS is to maximize for ROI. We want to make money, so we want to find the most profitable contests. But after building out our contest simulator, we discovered that this advice is just flat out wrong. As most of you have experienced, a 10% roi can play out in a multitude of ways, a multiple week losing streak followed by a big day, several close top 10s that don’t make the top 5, or a slow grind of cashes. These streaks of good or bad luck can be a wild ride, and as it turns out, if not managed correctly, can cause you to bust your bankroll entirely. But this variance can be managed, and if you use a proper contest allocation strategy, you can lower your variance and make a lot more money money in the long run.