The following is an excerpt from Dueling with Kings: High Stakes, Killer Sharks, and the Get-Rich Promise of Daily Fantasy Sports, Daniel Barbarisi’s book about a year spent attempting to become an elite daily fantasy player. The book is out now.
But it’s the work on DFS itself that defines these companies, and as Cal Spears says, that puzzle renews itself every day. In both DraftKings and FanDuel, it comes in the pricing of players and the setting, sizing, and pricing of contests.
Pricing players is fairly straightforward. The companies try to make sure that users can pick a few stars, but not build all-star teams; tough decisions should be a daily occurrence. Pricing is manipulated to encourage users to pick different players. Hot streaks drive prices up, cold patches drive them down. It’s all based on a computer algorithm, with human tweaks coming after—and the algorithm itself is fairly straightforward.
“When you say an algorithm, people think of Good Will Hunting shit on a wall. It’s not nearly as complicated,” a FanDuel staffer says.
That said, while not necessarily complex, player pricing is at the very foundation of the game and determines how skillful it is. A high-level DraftKings employee told me that they could make the game more or less skillful by adjusting player pricing; keeping the pricing tighter reduces the number of bargains, and that dulls the edge of the best players, who rely on those bargains to win.
Building the contests themselves is less straightforward, a more intricate and complicated daily dance. It’s also one that at times has created an arms race between DraftKings and FanDuel, with each one trying to force the other into money-losing situations.
It all comes down to what the industry calls overlay. According to UIGEA, the companies must lock in the sizing and pricing of tournaments before they begin. If larger, guaranteed contests don’t fill, the companies are on the hook for the difference. For instance, let’s say the companies announce a one hundred–person tournament with a $100 buy-in, and prizes of $5,000 to first, $3,000 to second, $1,000 to third, and they take the remainder as the “rake,” or the fees they use to make money. The problems start for the companies when the tournament attracts only eighty entrants. The same prizes must still be paid out. But now the company is losing money on this contest; it must give out $9,000 in prizes, after taking in only $8,000 in entry fees. That extra $1,000 is the overlay.
Build contests that are too large and you lose money to overlay. Size them too small and they will fill too early; you’ve left money on the table. Being a successful DFS company means finding the correct balance and predicting the demand that will exist on a given day. At FanDuel, that falls to the revenue team, composed of close to fifteen people, many of them math whizzes, and just as many former DFS players. That’s not surprising; working in these situations was alluring for potential employees. It made them better at understanding the concepts that would let them win when playing DFS on competitor sites. That led to the rise of players/employees like Boccio and Haskell, who then got the companies in trouble.
“One of the main draws of why the people we would hire were current—now former—players, was that they were like, ‘Oh wait, I can get a shitty paycheck, but my job is to help me do everything I want to do anyway.’ It attracted that. It wasn’t like we went out and looked for active players to work here. It was a natural fit,” says one FanDuel higher-up.
When building contests, FanDuel uses statistical modeling to look at similar days from past years, then projects current growth and user engagement from that baseline. They then analyze the actual matchups and upcoming games themselves to see if they will draw interest. Next, they weigh abnormal factors—holidays, for instance. The Fourth of July wreaks havoc with projection models, as do lesser holidays like Presidents’ Day or Thanksgiving. Valentine’s Day doesn’t actually change DFS engagement too much.
Finally, a human being will go through the computer’s results and make tweaks. FanDuel staffers show me screens where, through the day, red and green bars with percentage numbers track how quickly contests are filling, and project how close they will come to maxing out. If a contest isn’t going to fill, the company can either offer promotions, send reminder emails, or bump certain contests up in prominence on the site in order to encourage users to join. If a contest fills too early, they’ll often post another, a “special,” hoping to capitalize. But late specials almost never generate the proper demand and are often prime candidates for money-losing overlay themselves.
DFS companies sometimes plan for overlay to try to bring in users, a strategy DraftKings in particular pushes early in the NFL season to promote growth. But generally, they hate when their own contests overlay—and love when the other company suffers from it. And because they control so much of a limited market, the rivals can actually push one another into overlay conditions, by when and how they size their own contests.
Much of it comes down to when the contests are released; whoever releases their contests second can exploit either the aggressiveness or conservatism of their competitor. For a long time, FanDuel employees gripe, they believe DraftKings simply copied their contest sizing and pricing. FanDuel employees tell stories of waiting up all night to release that day’s contests as late as possible, forcing DraftKings to either wait indefinitely or post first. A DraftKings employee acknowledged that when they were trying to size contests, it was not uncommon for a company founder to enter the room, ask what FanDuel’s contest size was, and simply tell them to size their own contests larger.
It all culminated in what the companies call “Red Tuesday,” in the fall of 2015. It was September 14, the Tuesday after the first Sunday of NFL season, the biggest and most highly promoted day in DFS history. Both companies wanted to capitalize on the momentum and push new NFL users into also playing the final month of the baseball season.
“So, we were planning on launching a contest,” a FanDuel employee says. “DraftKings came out first, and they [sized it at] $1.5 million. It became the prisoner’s dilemma. Where you go, well, we could just size conservatively. And they’ll fill, and they’ll get their $1.5 million. Or, you could size ours at $1.5 million, and no one will make any money. We’ll both lose. And we chose the latter. And boy, does that data point stick out on the chart.”
The contests didn’t come close to filling. I remember it well—I told everyone I knew playing DFS to get entries in, and upped my own as well. Then I watched from Tropicana Field as the Yankees played the Tampa Bay Rays and I actually made a little money. Both DraftKings and FanDuel, however, lost hundreds of thousands on those contests alone. That, and the Haskell scandal, largely ended the overlay wars.
Post-Haskell, the companies have mostly been too busy trying to survive to fight silly battles over market share with one another. Instead, much of their energy is now put to use on two fronts: to address what I believe is well-deserved criticism, like that in Jay Caspian Kang’s widely circulated New York Times article, that their ecosystem is unfairly tilted to give the top players too much of an advantage; and to address what I see as less valid charges that the games are rigged, or not on the level.
To counter the first argument, both companies introduced widespread changes to the number of entries each player could have per contest, and across their sites as a whole. The main thrust of the criticism was that top players like Saahil were able to be everywhere at once: they could enter literally thousands of games on every level from $1 to $10,000 with virtually no restrictions, and they could also enter large-format tournaments and double-ups with hundreds of entries. [FanDuel CEO Nigel] Eccles said it was important to address these concerns, though he maintained that FanDuel was never in the business of trying to appease the sharks.
“Is there danger that the game is tilted towards the best players? Unfairly? That’s something we’ve spent a lot of time worrying about, and considering,” Eccles muses. “We built our business on casual players. We’ve always focused on being the most mass-market product. The accusation that we’ve been out trying to cater to the sharks is just completely untrue.”
But the reality has long been that any whiff of changes to level the playing field used to arouse the fury of the top pros. For a long time, the companies allowed these top players to hold sway, because their business was so important to keep; if one site blunted the pro edge, the pros and the massive revenue generated by their entry fees might run to the other site. As the sites grew, the biggest pros came to wield enormous power, and it’s not hard to see why. At his peak, Condia could often account for half of the total action on several smaller DFS sites, Saahil said. Piss him off, send him to your competitor, and it could be a death blow for your tiny but growing site.
There were other reasons as well to cater to the pros. People like me rail against them, but the harsh reality is that the Saahils and Condias of the world are exactly who generate the rake that keeps the prize pools as big as they are. For a long time, that’s what DFS users gravitated toward: the biggest prizes. It’s what fueled the rise of DraftKings, it’s what brought over the poker crowd, it’s what made regular people sit up and take notice when million-dollar payouts became a regular phenomenon in 2014 and 2015.
But there’s a tipping point there. As the high-volume pros became more and more effective and lethal, as more new players flooded into DFS, and as the companies became large enough that they were no longer as beholden to the big pros and their big fees, it became clear that something needed to change in order to preserve the ecosystem and protect the smaller players. But until the Haskell scandal, I believe the sites were too concerned with losing high-volume pros to really make big changes. Once the scandal occurred, the companies acted—they had to, as some inside the companies now admit.
“For a long time, [the pros] were so vocal,” says a top employee at one of the two major sites. “Now, we have much more of a pedestal to be like, come on. Come on. Before, we were still growing, and you might pander a little bit to it. Now it’s like, guys, look what’s happening here.”
The changes began with FanDuel raising the number of single-entry contests, which are seen as fairer outside the community, and introducing some limits on total number of entries. In late spring, FanDuel would go further, significantly limiting the number of head-to-head games players could enter.
DraftKings introduced a number of three-entry-max contests— which I love—and limited tournament multi-entry sizes to 3 percent of the total field. They later introduced many more restrictions that went a long way toward making the contests much fairer.
Both companies have long trumpeted the fact that they offer beginner contests restricted only to new players. But to me, beginner contests are mostly a joke and don’t go nearly far enough. Both companies long offered fifty beginner games per sport; those could be wiped out on the first night, following an otherwise intelligent strategy of entering twenty $1 and twenty $2 head-to-heads, and a few $3 tournaments. Suddenly it’s day two, you’ve invested $50, and you’re now forced to swim with the sharks. But beyond that, I’ve found that the changes the sites have made are sincere efforts to create a more balanced game, and they’ve made a significant difference.
The other charge they needed to fight was the contention that the games were literally not on the level, that either there was “insider trading” going on by employees of the companies or that players were using computer programs—commonly called scripting—to skirt or ignore the rules and gain an unfair edge.
They took care of the first issue relatively easily—though far too late—by banning employee play on competitor sites. Usernames and identities of DFS employees are now circulated among the sites. And if employees are found to be using opponent sites or playing through relatives’ or friends’ accounts, they’re out of a job.
Eccles looks at his failure to ban employee play as one of his biggest mistakes—it was something he simply didn’t think about until it was too late. It was part of an era when the companies were growing so fast that they hadn’t realized they needed to adopt entirely new sets of standards. He now looks back on the hilarious HBO piece featuring Seth Rogen and John Oliver as a wake-up call.
“When John Oliver did the takedown on us, my reaction was, wait a minute, that’s really unfair,” Eccles says. “They do takedowns of big companies, like FIFA. The people who can’t defend themselves. And then I thought, oh, wait a second—we are those guys. That’s the thing. We thought we were the little guys. And then I realized, we’re not. We’re just not the little guys anymore. We raised $300 million. People don’t think of us like that. To be honest, when I watched that show, that was the point where I was like, okay, that’s where we went wrong.
“So yes, absolutely, I would like to go back, and there’s a point where we could have spotted it. There should have been a point where we said, look, we’re now a really big business, we should ban employees from playing on any site,” Eccles says.
Adapted from Dueling with Kings by Daniel Barbarisi. Copyright © 2017 by Daniel Barbarisi Reprinted by permission of Touchstone, a division of Simon & Schuster, Inc.