FanDuel Strategy: How We Doubled Our Money In 2014

September 12, 2015 - by David Hess

As we get started with our 2015 daily fantasy football analysis, we wanted to recap what we learned over the course of last year’s successful 8-week experiment playing daily fantasy football on FanDuel for the first time, which ended up netting us a 100%+ return on capital.

We’re going to be spending more time playing this year, so the goal is to start with what worked for us last year, and build off what we’ve learned over the course of the 2015-16 NFL campaign.

With more playing experience, we’ll also should get more insight into whether our success last year was due more to skill, or to luck. We hope the former. 🙂

Of course, what’s harder to predict is whether the dynamics of playing on FanDuel will change at all this NFL season. In short, will the average player be easier or harder to beat than last year? We’ll find out soon enough.

As for what worked last season, we split our accumulated learning into two categories, based on the format of game.

Head-to-Head & 50/50 Strategy

In these games, you’re either playing one other opponent head-to-head, or participating in a bigger contest where close to 50% of players win prizes.

The general strategy theory is that you don’t want to get too risky with your drafting. Getting a pretty high score consistently (e.g. a “high floor”) will win you more money in the long run than a boom-or-bust oriented strategy that nets the occasional fantastic score (i.e. a “high ceiling”), but also a bunch of below average scores.

A discussion of the historical data that forms the basis of this strategy can be found in last year’s post on FanDuel scoring by position, but the basics of my current approach are:

  1. Pay for stud QB and RBs. These players generally have high ceilings, partly because they touch the ball so often. It’s very rare to pay for, say, Tom Brady and end up with single digit points.
  2. If it works out, try to get RBs who are playing against each other, in the hopes that at least one of the two teams will jump out to a lead and pound the rock all game. The performance of the two opposing RBs should be inversely correlated, decreasing the chance I get two busts at the position.
  3. Don’t spend too much on any one WR, if it prevents me from getting other players I want. Wide receivers play a high variance position. When you only get thrown to a handful of times a game, even the best receivers have a chance of have a low-scoring game. So when choosing between splurging on a WR or splurging on a RB, go with the RB.
  4. Be willing to spend on DEF and TE if there’s a matchup that looks really juicy. There are only a handful of games either week where a tight end is actually likely to have a decent scoring game. The majority of tight ends don’t provide much value. So it can make sense to spend a bit of money to lock in some production at the position.
  5. Don’t spend money on K. They’re too unpredictable. Basically, they are high variance and low upside. At least at wide receiver, you get high variance and high upside.

Again, the general idea here is that for head-to-head and 50/50 contests, you want a low-variance team. You’d much rather have a 100% chance to score 15 points, than a 50% chance to score 20 and a 50% chance to score 10. Both of those scenarios lead to an expectation of 15 points scored, but the former will win more 50/50s.

Tournament & GPP Strategy

On the other hand, when entering a tournament with a guaranteed prize pool where only the top few percent of finishers get paid, you want to emphasize high-upside rosters. That idea led to the following basic strategy, which we originally laid out in our 2014 Week 7 FanDuel Strategy post:

  1. Identify several teams where the QB has the potential to have a monster game. That list should include either stud QBs (Manning, Luck, etc), or QBs facing a terrible pass defense. Average projected value per dollar here is less important than the possibility of a FanDuel point scoring explosion.
  2. For each QB, pick 2 players to pair with the QB. Projected value per dollar can play a bit more of a role here, but you still want high upside players.
  3. Let RotoWire optimize the rest of the lineup … with some manual tweaks. Basically, I enter each stack I compile via Steps 1 & 2 above into the RotoWire lineup optimizer. When I get the results, I then check them against numberFire’s projections and against team defense-against-position stats. That begins a cycle of manual iterations to finalize the lineup: toss players that those other two sources weren’t very high on, re-optimize, and repeat until you’ve got an optimized/vetted team.

The hope here is that among 5 or 6 stacks that we identify with this QB-oriented approach, 1 or 2 will hit big and snag larger prizes that more than make up for the entry fees for the bigger number of stacks that will whiff. This is the Sammy Sosa strategy; lots of strikeouts are just fine as long as we hit enough dingers.

It should be noted that there are lots of strategies floating around out there regarding how to play different types of tournaments. We’re not saying our approaches are the best, we haven’t been using them for very long, and we’ve haven’t had the opportunity to back test them at all. But the early results using them in our 8-week experiment last year looked promising enough.