Here at RotoFanatic, Daily Fantasy coverage will serve as one of the pillars of our 2020 content campaign. It’s bound to be a wild ride with such a short season and frustrating usage patterns. In particular, picking pitchers is going to be a nightmare.
While we’ll surely have plenty of analysis about the state of 2020 DFS, today’s subject is meant to be rather evergreen. I’d like to discuss my personal approach to playing – and enjoying – the game.
I’ve broken this post into five sections. I begin by discussing the point of daily fantasy sports. Then we’ll touch on some very daunting math after which I’ll explain how I have fun and manage my bankroll. Lastly, I’ll give some basic tips and tricks for eking out monetary advantages. You can use these links to skip around:
- Redefine Your Purpose
- Why Is It So Hard To Make Money?
- How I Play
- Bankroll Management
- Beating The System
First a couple of credentials and whatnot. Since 2014, I’ve written around two thousand articles about daily fantasy. I’ve also played every season – with mixed success. The Wild West of 2014 was a hugely profitable time as the competition either used shady batter vs. pitcher data (BvP) or inefficient algorithms.
Between 2014 and 2015, the sharps sharpened and BvP (mostly) went out of vogue. The game got harder. I can give this very qualified brag: excluding site rakes*, I have made money in every season. In other words, I outperformed my human and algorithmical opponents, but the house still took their pound of flesh.
*A rake is the cut of each entry taken by a site. More on these in a moment.
Over the last half-decade, I’ve developed a distinct approach to daily fantasy. I believe it to be a sane and rational response to the rising tide of algorithms and lineup generators.
Redefine Your Purpose
As with all gambling (and yes, DFS is gambling), it’s tempting to view the game as a way to make money. It is possible to make a living playing daily fantasy. We know this because some people have done it. If you’re looking to DFS as a source of income, first ask yourself these questions.
- Am I using a proprietary algorithm?
- Does it include unique data in such a way that it can be expected to outperform other algorithm users?
- Have I mastered the lineup theory for large multi-entry guaranteed contests (GPPs)?
That’s three no’s for me. How about you?
Put extremely generally, there are three ways to profit in any meaningful way: spike large GPP wins, grind a narrow edge in high-volume 50/50 (or similar style) contests, or play small, high-stakes games.
For the most successful DFSers, volume is the secret sauce. Below I’ve posted a sample KBO contest from a major DFS site. User TLorz has opted to submit 51 entries. If done properly, these will overlap in such a way as to maximize the chance of winning the contest. It’s difficult math, and, frankly, I’ve never successfully solved for X. If you don’t know exactly what you’re doing, sending forth an army of entries is a wonderful way to lose money really fast.
If you insist on giving multi-entry a go but haven’t built your own system, there are dozens of lineup generators out in the wild. I don’t use any of them for this purpose, so I can’t give a recommendation.
Let’s be honest, you’re probably not reading this article if you’re a professional DFSer. So, if we’re not here purely to earn an income, then why play? For me, it’s about enjoyment, competition, and the rush of spiking a rare GPP victory. Just like playing blackjack at a casino, I’m trying to win while being realistic about my odds. And I’ve styled my process to be fun for me.
Why is it so hard to make money?
There are two reasons why it’s so hard to make money playing DFS – the competition and the rake. Ultra-deep, low-stakes GPPs do have a soft underbelly of lottery chasers entering non-competitive, long shot lineups. Even so, the level of competition across all contest-types is such that you really need to know what you’re doing to have a positive expected value. Let’s use some very simple examples.
Say you’re in a 10-entry, $10 contest. The winner gets $100. If everybody is equally skilled, your $10 buy-in has an expected value to you of $10. Let’s say you’re more talented than your rivals. You have a 12.5 percent chance to win the $100. This implies you’re roughly 25% more talented than your opponents. You have an expected value of $12.50 on your entry. Imagine how many times you have to run this league to earn an income! Play this contest 1,000 times and you’re expected to win a whopping $2,500. Oh! Here comes the taxman to collect his 30 percent. Now you have $1,750.
Now let’s add a layer of reality to our simple example. It’s a multi-entry tournament, and you are an ordinary entrant. One of the participants is a ringer. They have a $13 EV per entry, and they enter three lineups. They expect to win $39. That leaves everyone else with around $8.70 expected value per entry. If you are a non-algorithm user playing public GPPs, this resembles your reality.
Here’s where the math gets out of hand. The platforms have to take their cut. The states where they’re allowed basically extort them so it’s an extremely expensive business to operate. That’s why you might find rakes as high as 16 percent of a pot. In the above pictured KBO contest, the site is pocketing $4,400, a 14.9 percent rake. Ready for one more example?
Imagine the same 10 entry, $10 contest. Everybody has an even chance to win. The prize pool is $85 because the site took a 15% rake. Your expected value per $10 entry is $8.50. Now add in the scenario with the ringer. Say their EV is $11 per entry and they enter three times. This implies you have an expected value of around $7.40 per entry.
This has serious and obvious implications. To earn an income from DFS, you need to be much better than the competition. You must outperform your rivals AND the rake. You also need to succeed at a high volume. To do all this predictably and without a razor-sharp algorithm is, in a word, impossible.
How I Play
To simplify, I have one operating principle – I play to have fun. To paraphrase a character from a favorite story of mine, you have to be realistic*. Winning is fun. By the transitive property, I play to win.
*bonus points if you can name the character.
As we’ve already covered, the deck is stacked against us. Since I’m not using an algorithm and I don’t have a strong grasp on multi-lineup design, I only enter one or two unique lineups each day. I use a basic, flexible process for building them. I do my best to pull in information the public might overlook or mistakenly evaluate. And then, to top it off, I keep it cheap.
This winds up looking almost embarrassingly unsophisticated compared with what’s out there in the marketplace. Nonetheless, for me, it’s enjoyable, sufficiently adjustable to my whims, and successful enough to prevent me from losing money.
Mid-2019, I developed a new process, an adjustment to the homertastic meta. The foundation of my lineup was formed by chasing value home runs. I had daily projected home runs from SaberSim (a DFS projection system), and I had dollar values. I simply divided the two together to find the best home run picks, ignoring all other traits.
In most normal slates, those formed the backbone of my lineup. Then I approached pitching, an always complicated endeavor. Some days, pitchers are quite dear. Others, you can get away with value plays. That would then decide if the remainder of my offense was filled with chalky studs or deep value plays. I give a little more detail about this process here.
A core tenet to how I play DFS is to use scouting and unusual stats. The purpose is to build competitive lineups with a slightly different look than those presented by the many and often overlapping projection systems. Especially in large contests, it’s tempting to be contrarian with your picks so you have a unique lineup; however, contrarianism for its own sake is a losing strategy.
While hardly revolutionary, this often simplifies to chasing hitters with traits aligned to the opposing starting pitcher. We all know about lefty and righty platoon splits, a mainstay of the DFS advice column industry. Did you know fly ball and ground ball hitters often have splits too? From the hitter’s perspective, you want opposing traits such as a fly ball hitter facing a ground ball pitcher. For pitchers, it’s great when you can send a fly ball arm against a collection of fly ball hitters.
Another common pivot I’ll use is to target exceptionally bad bullpens when that team is using a decent starter. Attacking Matt Boyd was a fairly successful strategy, especially when he faded in the second half.
Building lineups is fun, but it’s also important to use them in the right way. I focused on the FantasyDraft platform last season for several reasons – the biggest of which is it’s easier for “normals” like you and me to win money. Their prize pools aren’t enticing enough for the professional DFSers, but they’re also big enough to hold my attention. If DraftKings and FanDuel are the National and American Leagues, FantasyDraft is the Pacific Coast League.
If you’re playing on the big sites, you can improve your odds by focusing on single-entry contests. The lower-cost contests (i.e. $10 and under) will mostly be filled with normal players. I’ll also throw a single entry into a large, low-cost GPP just in case I spike the nuts. It’s a negative-expected value play, but it’s not too hard to justify a $2.75 EV on a $3 entry if you’re careful elsewhere. Posting a token lineup in huge GPPs also keeps the dream of a life changing win alive. It’s a lottery – and the odds are better than any Powerball or Mega Millions.
The most important part of enjoying and responsibly playing DFS: Manage Your Bankroll.
Honestly, this could be a standalone article, but I’ll stick to hitting some basics. You can also google poker guides like this one to get more ideas.
Just like when you sit down at a blackjack table, you should know how much you’re willing to lose. That amount should inform the maximum you spend on DFS contests on a given day. As you win or lose, you can adjust your entries to match your new bankroll.
I’ve seen it suggested that you should have anywhere from 10 to 50 days of buy-ins in your bankroll. In my opinion, it’s just a matter of preference. Let’s say you have a $500 initial bankroll and you want 20 days of buy-ins. That implies you can comfortably spend around $25 per day.
If you spike a $1,000 victory ($1,500 bankroll) you can either grow your daily entries to $75 or stay at the same $25 level. Heavy losses are most common when trying to jump from small-cost to mid- and high-cost contests. If you’re a conservative gambler, stick with where you’ve had success.
Conversely, should you zero out your first 10 days ($250 bankroll remaining), you might want to cut your daily buy-ins down to between $10 and $15. This will extend the life of your bankroll even if it also serves to limit your upside.
Ultimately, you have a lot of flexibility with how you approach bankroll management. You can tailor it to complement your risk tolerance.
Personally, I focus almost exclusively on low-cost contests. My daily writing commitments during the season are pretty steep. I often have only a few minutes to customize my own lineups and later verify weather, umpire, and other conditions. Since I have to rush on some days, unforced errors are not unknown. As I said before, you have to be realistic.
I typically start a season by playing only $5-and-under games for a total of around $20-$25 per day. My initial bankroll is usually only around $300. On one occasion, I zeroed out in April! I switched to $3 of daily entries and slowly rebuilt my cash reserves. Those micro-cost contests are also the easiest to win. By the end of the year, I had made my way back to $20 per day and even made a tiny profit. It was a rewarding experience.
I’ll occasionally make undisciplined forays into large contests if I have a strong scouting find (I made a lot of money on Aristides Aquino last year before everybody picked up on him) or if I’m chasing a big overlay (more on those in the next section). Treat these pivots out of your usual bankroll rules as an advanced play. In other words, resist the urge to ramp up too quickly until you’ve mastered your target contests. To be honest, I’ve lost money on these so-called advanced plays.
Beating the System
I have three tips for beating the system. The first is to use a daily projection system. The most ambitious among you can build something from scratch. There are also plenty of these on the market. I’m most comfortable with the projections underpinning SaberSim (Steamer with some daily adjustments) and Derek Carty’s THE BAT. Again, I cannot comment on their multi-entry lineup generators. Both are paid services stuffed with features.
Some contests have low or no rake. These are usually associated with a website. For example, Baseball Prospectus ran a no-rake GPP on DraftKings during the 2017 season. Through my The Daily Grind column on FanGraphs, I ran a similar contest on FantasyDraft for the 2018 and 2019 seasons. With little or no rake, it’s easier to either luck or skill your way to the victory lane. As of mid-2019, FantasyDraft has done away with rakes, instead replacing them with a subscription fee. If you’re in the industry like me and have freelance income, the fee has positive tax implications.
If you thirst for the occasional big money play, be on the lookout for overlays. An overlay occurs when the prize pool of a GPP exceeds the buy-in of the entrants. For example, using the KBO example above, an overlay will occur if there are fewer than 1,666 entrants. Overlays tend to be rare on the major sites because everybody is hunting for them. Keep an eye out for weird conditions like uncertain weather to drive down entries. Second-tier sites are more likely to have overlays, but they’ll also have smaller prize pools.