How to Manage Your DFS Bankroll

An in depth look at the optimal ways to build a DFS bankroll without increasing your risk

The goal of standard fantasy leagues is to win the championship.  On the other hand, the goal of Daily Fantasy Sports (DFS) is to increase your bankroll as much as possible.  However, in my last article about the different types of games we realized that the games that offered the highest rate of return (GPP’s) also offered the highest risk.  Comparing your rate of return to the level of risk on the return is one of the fundamental rules of financial investments and it is no different in DFS.

Here’s an example of why bankroll management is so important.  Let’s assume we are the best DFS GPP player in the world.  We can win large GPP’s that include 1,000 users at a rate of one every 100 games.  If we had an unlimited amount of money and we could start a brand new GPP tournament every minute this would give us almost a 100% return on our investment (less site rake).  Unfortunately we don’t have unlimited money or an unlimited number of games available each week.  We are constrained by the number of games in a year and by our bankroll.

Now let’s assume our bankroll is $100 and each tournament costs $10 to play.  If we simply entered the next 10 tournaments and expected to grow our bankroll to $200 we would be sorely disappointed most of the time.  In fact over 90% of the time we would be broke by the end of the 10 tournaments.  This is despite winning at such an incredible rate! In other cases we may have gotten very lucky and won $20,000! Again our goal is to have the largest bankroll possible by the end of the year and this means we can’t lose all our money.  To do this we have to manage our risk.

Luckily for us there is a formula called the Kelly Criterion which is used for this type of risk management.  It tells us exactly how much of our bankroll we can invest in any specific game by taking into consideration the payoff odds.

The Kelly Criterion formula is a simple formula and you can find several calculators on the internet to easily calculate it. It is defined as:

                F = (bp-q)/b

  • F is the fraction of the current bankroll to wager;
  • b is the net odds received on the wager ("b to 1"); that is, you could win $b (and get a return of your $1 wagered) for a $1 bet
  • p is the probability of winning;
  • q is the probability of losing, which is 1 − p.

All of the inputs into the formula are very easy to calculate except for the probability of winning which unfortunately is a very large assumption.  On top of that it is very hard to determine.  We won’t know this true winning percentage until we have played thousands of games.  I highly recommend if you are a new player to continually track your winning percentage and all current players to look at your historical results. 

The quickest way to get to this win rate is by playing 50/50’s since you are playing so many people at one time and you don’t have to pay for each of those games.  Keep in mind your true odds are always going to be changing as we improve as are the industrial baselines so you should always continue to monitor it. But even without knowing an exact win rate the formula can help us compare between the different game types.  All things equal be conservative in estimating your actual win rate.  It’s better to estimate low and not win enough than to overestimate and go broke.

So what percentage of our bankroll should we wager based on some sample winning percentages in 50/50’s with the assumption that a $10 entry will pay out $18.   

Odds of Winning Bankroll to Wager
50% 0%
55% 5%
60% 15%
65% 26%

The first thing that is important to note is that if we are only an average player and are expected to win 50% of the time we shouldn’t be playing DFS.  This is because over time the rake would take all of your money. But as you increase your odds of winning you can play more and more of your bankroll.  But even if you are expected to win 65% of the time you should still only be wagering 26% of your bankroll on any lineup. 

For illustrative purposes let’s assume that our odds of winning are between 55% to 60% which means we should be playing 10% of our bankroll per the formula.  But just because the formula says we should only play 10% of our bankroll that doesn’t mean that’s all we can play in a given week.  Remember there are several different types of games we can play and Kelly Criterion is designed for games where we simply win or lose.  This is exactly the type of game that 50/50’s are. If you entered 1,000 50/50’s with the same lineup you would most likely win all of them or lose all of them unless your lineup is exactly in the middle then you will split some. 

However, in head to head games this isn’t the case.  It is true that we either win or lose every matchup. However, if we play 1,000 games and finish in the top 80% we will win close to 800 of the games and lose close to 200.  This isn’t a simple win or lose situation. This makes our risk a lot lower and we can play more of these games.  In this case I think we can actually play approximately double what the formula tells us which in this case would be about 20% for our top head to head lineup.

Notice I mentioned this is what we should use for our top lineup.  If we actually play multiple lineups we can again increase the amount of our bankroll that is in play.  How much we can increase our in play money depends on how correlated it is to our first lineup. This means if none of the players in our second lineup are the same as our first lineup we can play that lineup as a fully unique lineup following the same calculations as above. 

However, we also have to assume that our second lineup won’t win as much as our first lineup, especially the more uncorrelated it is because the best values will always be in our first lineup. Let’s assume our second lineup has a 56% chance of winning. In this case we can play this lineup with 7% of our bankroll, or 14% in head to head games.

On the other hand let’s assume 60% of our money spent on our number 2 lineup is the same as our first lineup. In this case we can’t use the full fraction the formula tells us because we are putting too much weight on the same players.  I think a fair estimate for this lineup would be to decrease the amount we spend on this lineup by about 60% of what the formula tells us.

Think about an extreme “second” lineup that was exactly the same as the first lineup.  Obviously these two lineups would be 100% correlated and we wouldn’t want to play this lineup again.  We can help ensure we are playing different players and keeping our lineups uncorrelated without significantly decreasing our odds of winning by playing some of the smaller game types like early games only for some of our secondary games.

Let’s put this plan in action for a sample week with the following assumptions:

  • Bankroll = $1,000
  • We like three lineups and project the following odds of winning
    • Lineup #1 = 59%
    • Lineup #2 = 58%
    • Lineup #3 = 57%
    • Each of the lineups have unique cost of players as follows:
      • Lineup 2 = 70%
      • Lineup 3 = 50%
      • Allocation of 80% on head to head and 20% on 50/50.

In this case we could wager amounts that are in line with the chart below. But first let’s walk through the calculation for lineup 2 as an example.  We expect our odds of winning to be 58% and we will receive $18 for every $10 bet.  This means we should bet about 11% of our money on this wager.  However, we noted that only 70% of our money spent on players are unique to lineup 1 so our adjusted wager will be 8% for 50/50’s and 16% for head to head.  Finally, we split our bankroll to the different games keeping in mind we wanted to wager 80% on head to head and 20% on 50/50’s. This allows us to spend 80% of the 16% we calculated above, or 12%, on head to head and 1.5% on 50/50’s.

Lineup #

Head to Head

50/50

1

$208

$26

2

$123

$15

3

$72

$9

Total

$420

$53

This method suggests we should wager approximately 47% of our bankroll on these three lineups.  On average I think anywhere from 30 to 50% of our bankroll can be in play at a time, but I wouldn’t risk going much above that. From here we could split these amounts up even more if we wanted to.  For example, we could play half of lineup #1 on Thursday night and half on Sunday.  Lineup #2 might be just for primetime games, etc.

As I previously mentioned I want exposure to these other games as they have various advantages we can exploit and allow us to diversify our players.  Unfortunately there isn’t nearly as much liquidity in these games so I don’t think they should command as much of your money as the standard games but I would recommend getting my money split up something like this:

  • Sunday – Monday: 60%
  • Thursday – Monday: 20%
  • Various mini games 20%

There are numerous ways we could split these lineups up.  However, one thing I would be very hesitant to do is play too many different lineups with the same players in the same game category.  The reason for this is some sites don’t let you mass change whole lineups, instead they only allow you to change individual players. This makes it tough when you decide that you want to remove a player from one lineup but not from the other one. 

Last but not least, we want to remember that we should be setting aside about 3% to 10% of our bankroll for the GPP’s that are going to offer us our highest returns.  We went through all of this analysis on how to manage our risk so we could play some more of the high upside game after all.   I tend to not use the Kelly Criterion for GPP analysis because it isn’t designed to calculate these types of games at all.  I feel like these wager amounts are a good combination of risk and reward but certainly use what you feel comfortable with.  If you are on the high end of this estimate you may want to limit your exposure to 50/50’s when determining how you are going to split your money between 50/50 and head to head since they are the higher risk cash game. If you are on the low end you can increase your exposure to 50/50’s.

This is just one approach to managing your bankroll in DFS.  It is a pretty aggressive approach too since the Kelly Criterion formula is derived for an optimal amount of aggression.  If you don’t feel comfortable wagering these percentages of your bankroll don’t be afraid to scale them back and always be conservative with your odds of winning. 

Winning in DFS means having the best return on your investment over the course of your playing career.  Bankroll management is the only way you can accomplish this goal.  Even if you are the best DFS player in the world you will eventually go broke if you wager unwisely. But if you manage your risk it will go a long way towards helping you reach this goal. 


More articles from Steve Buzzard

See all

More articles on: Daily FF

See all