Connect with us

Diversification

Game Plan: DFS Diversification Primer

Published

on

Finance types often find success as DFS players, in large part because many concepts from their field transition well to DFS. In finance, diversification refers to the concept of spreading your assets among a variety of investments to ensure you’re not overly exposed to negative outcomes, allowing you to invest more capital safely. In DFS diversification, the principle remains the same. 

Lineup outcomes, similar to investment outcomes, are volatile, unpredictable, or—the term of choice in DFS circles—“high variance.” All three of these terms, though their definitions vary, refer to a similar concept: how the outcomes differ from a reasonable prediction.

Any fan knows that sports are unpredictable, so attempting to achieve a high degree of accuracy in a single slate is futile. While it’s fun to sweat a contest, the results from one night are not particularly useful information whether your lock of the day busts or you win the whole prize pool! To get a true sense of your skill and profitability you will need a large amount of data collected over many slates of contests. DFS Diversification can help in this regard.

Rule #1: More lineups are better than fewer

One way to play more lineups is to reduce the entry fees of your contests. For example, if you are playing $20, enter 80 lineups at the 25 cent buy-in level rather than a single $20 entry. This has the added bonus of not only reducing variance but also increasing your return on investment because, as noted in the Contest Selection Primer, smaller entry fee contests have weaker competition.

Another means of increasing the number of lineups you play without increasing entry fees is by playing unique lineups in every contest. Too many DFS players fall in love with their “optimal” lineup and recycle them in all contests on a slate. It isn’t single-entry players alone who fall victim to this thinking, most mass-multi-enterers use the same 150 lineups in every large-field GPP. 

Given the high levels of variance, the difference between the top-projected lineup and the second-projected lineup or even the top-150 lineups and the second-150 lineups is insignificant. And remember that top-projected lineups aren’t always “optimal” in terms of profitability when accounting for game theory.

Rule #2: Play more sports and slates

Contests from different slates on the same night should be treated as independent of one another. I have found that even two game slates can be beatable over the long run, so I recommend playing every slate to reach your long-term goals faster. Since the results of each of these slates are uncorrelated, you can invest as much as you do on the main slate the same night without taking on much more risk than waiting until the next day. The same definitely applies to different sports: if your bankroll strategy is investing 5-10% of your bankroll on a given day, you can double that by playing a second sport.

Rule #3: Reduce inter-lineup correlation

In addition to spreading out your entry fees over a large number of lineups, you must ensure that your lineups have significant differences or else their performance will be strongly correlated. The issue with having lineups that are too similar is that your entries will all be clumped together, increasing your volatility, and decreasing your expected profit because if the narrow outcomes you hope for happen, you’ll frequently be competing against yourself for the top positions. To reduce your inter-lineup correlation, you can either build your lineups by hand or you must master a lineup optimizer! The optimizer I use is Fantasy Cruncher, so I’m going to focus on features they have to reduce your correlation between lineups.

Unique Players 

The unique players function offered by most optimizers specifies the number of overlapping players any two lineups can have. For example, if you specify three unique players then every lineup you generate will have a minimum of three players that are different from every other previously produced lineup. This is an effective method of reducing inter-lineup correlation as it ensures that each lineup will be significantly different from the rest. The ideal setting for unique players will vary based on the possible number of lineup combinations but I would recommend somewhere within the range of two to five, increasing it with the size of the slate.

Maximum Exposures

The maximum exposure setting removes a player from the player pool available for the next produced lineup as soon as that player goes over his maximum exposure limitation. Maximum exposures can be applied to your Fantasy Cruncher lineups either through the “each” or “group” method. I recommend using the Each method because the Group method will create lineups that are either very chalky or contrarian, and I like to have a mix of both types of players in my lineups.

With Each, the players will be removed from or added back to the player pool whenever their exposure in the previously created lineups exceed or dip under their limit. I recommend setting a different exposure for every player to ensure maximum differentiation between your lineups. Some of the lineups created by this method aren’t strong though, so I always sort by projection and eliminate the bottom 25%.

Finding the right exposure limit for every player is the challenge. You should consider the following factors: projection, salary, volatility, and slate size. You should strive to come up with an approach incorporating all of these factors, but an exposure scheme isn’t a simple mathematical problem as its success relies on complex game theory components. The only way to know if it works will be to test it out. Start off slow when testing a new strategy and only ramp up your volume after achieving some success.

Randomness

The “randomness” feature essentially applies a different random factor to each of your players’ projections for each lineup you generate. In effect, you are using a slightly different set of projections for every lineup you create. Because small differences in projection can lead to drastically different lineups yet have only a marginal effect on your ability to cash in GPPs, this can be an effective means of achieving diversification. The ideal setting will depend on the variance of the sport and the slate size, but I’d recommend something in the 2-5% range. This isn’t a method I personally use, but I’ve heard it can be utilized effectively.

Having a good DFS diversification strategy will help you have more consistent results and stay in the game. Using diverse lineups on a given night will increase the chance that you have a good day but decrease the chance of a killer one, which is a trade-off I like to take. Entering multiple slates will allow you to increase your exposure to fantasy contests each day without increasing your risk. Finally, having differentiation between your lineups will allow you to have your results spread out more throughout the tournament so you have less volatile results and decrease the chance of beating yourself out for the top spots. Employing these strategies is a key to being a successful DFS player!

A self-proclaimed U.S. history buff, stats enthusiast, and rabid fan of anything Detroit (yes, even the Lions), Awesemo started playing daily fantasy sports in 2013 and never looked back. After graduating with a degree in mathematics from Washington University in St. Louis in 2008, he began developing advanced statistical models and applied his unique skill-set first to poker, then as a full-time DFS player. He has since dominated DFS, becoming a fixture on leaderboards and cementing his position as the top-ranked overall player according to RotoGrinders. Competing in most guaranteed prize pool tournaments, he has ranked in the top 3 across major sports, including NBA, NFL, NHL, PGA, and MLB. He enjoys sharing his knowledge of fantasy sports with new and experienced players alike. Career Highlights: #1 Overall Ranking; 14x FanDuel $100k winner, 11x DraftKings Hundred-Thousandaire; 20x Live Final qualifier. You can contact him on Twitter @awesemodfs or by emailing support@awesemo.com.

Continue Reading
8 Comments

8 Comments

  1. M D

    February 16, 2018 at 9:17 pm

    Great read. But wouldn’t your prefer to win big by taking a higher risk/ less diversification than just cashing every night?

  2. ROBERT SHIVERS

    February 17, 2018 at 12:09 am

    Great read i never really understood how to use the optimizer

  3. M D

    February 17, 2018 at 7:54 am

    I think to talk about how you still use a tight core in nba on decent size slates you are normally 85%-60% on 4 or 5 main guys in the nba regardless of what you play in whether it be 20 entry or 150 entry you still send out the same core guys from my research over the last month of your lineups. Then you must use the optimizer to set unique players.

    • cjh13

      October 27, 2018 at 9:34 am

      Question, What if you set the unique to groups instead and create a rule like must use at least 3 of these guys? I have done my rewind and I am getting closer and closer to beating the top score. I get rewind crunches now in the Actual score of 330 … My pool is always spot on as well, It just seems like I can not create enough lineups it always spits back like 1 or 2 or sometimes 10 but if only I knew how to apply my method and rules with just my tight core and spit out a whole bunch of lineups! Id be golden – Thanks In Advance

  4. Cedric Payne

    February 21, 2018 at 1:06 am

    Very Helpful Read

  5. Steve Huckaby

    June 15, 2018 at 12:10 am

    that is amazing

    • BGV21

      July 5, 2019 at 11:34 am

      What would be the recommended randomness for mlb?

  6. aus10

    August 11, 2019 at 5:51 pm

    How do you determine how many unique players to use given a certain slate size?

You must be logged in to post a comment Login

Leave a Reply

Latest Video