How to Use a DFS Lineup Optimizer to Take Down GPP Contests (NBA)


Using a DraftKings/Fanduel DFS Lineup Optimizer?

A question I get a lot over here at Analytic City is "How do I use your lineup optimizer successfully." Now, I'm not claiming to be an expert on this by any means, but as the builder of many DFS lineup optimizers and a longtime DFS player with some decent success (and a few takedowns!), I will share a few points I think are critical to consider when using our (or any) lineup optimizer. Note, this post assumes you are using somewhat accurate projections (as you would find on any credible site).

1. Start by looking at the raw optimal lineup. I always use the pure optimal lineup as a building block for the night. Three main reasons for this, Total projected ownership, total projected points, and potential plays I may not have considered.

The Total Projected Ownership* of the lineup is the most important. If I look at the pure optimal lineup on a large slate, then make some adjustments/locks, and look at my new lineup only to see the total ownership has gone up by 75%, I know I probably need to reconsider some of my adjustments/locks. Hypothetically I have not only just created a less optimal lineup (fewer projected points), but also a more highly owned lineup. Of course, given that these are only projections and not guaranteed to be accurate, there is some margin for error, say 50% depending on the slate, but using the initial pure optimal lineup as a starting point can give you a sense of where your lineups final ownership should end up.

*Note: If you're still not sure why ownership matters so much in DFS, give this example by FantasyLabs a read. We are not affiliated with FantasyLabs in any way, I simply think they do a nice job explaining this.

This process is similar for total projected points. I do not want to create a lineup that projects for significantly fewer points than the pure optimal lineup. If I do, then I know I've gone too far with my adjustments/locks.

Finally, I like to look at the pure optimal just in case there may be a play or two that I wouldn't otherwise consider.

2. Do research, and find high upside, low ownership plays. The nice thing about an optimizer is you don't typically have to lock in the obvious chalk plays, the optimization process will do that for you. Instead, spend time looking for players with huge upside and low ownership. You can do this by looking at statistics like our potential rating or advanced matchup statistics. Or by just thinking critically about whether or not a player truly has potential to 6-7X value on a given night. I don't think there's much more to be said for NBA. Perhaps you could argue that stacking stars from opposing teams in the same game is advantageous in the long run, but unlike other DFS sports, game stacking is far less important in NBA DFS. I typically find 2-3 of these players each night before running any sort of initial optimization.

3. Finally, exclude low ceiling players. You just ran the optimizer and noticed that Rudy Gay was in your lineup at $4,800 tonight? Given the fact that he's only played 20 minutes in each of his last 2 games it'd be hard to envision him going for the 30-35 DK points he would need to around 6-8X value (what's needed to take down a GPP most nights). In cases like these, go ahead and exclude Rudy in favor of a player with higher upside.

You may be thinking to yourself that a lot of this is subjective: I would first argue there are ways to quantify potential/upside: (We do this through out AC Potential Ratings, which look at the players salary, projected minutes, standard deviation of scores this year, and maximum three scores to create a rating 1-100 of the 'potential' each player has on any given night.) I would also argue that it's less subjective than you think. Looking at players past performance, matchups, projected minutes, historical minutes and more can give good insight into their upside on any given night.

The real power of any lineup optimizer is the ability to show you different builds and combinations of players you would never be able to replicate when hand building lineups. Couple that with some research, critical thinking, and ownership projections and you put yourself in a position each and every night where you can build a lineup that has Takedown Potential!

Hope you enjoyed! If you're interested in using our NBA DFS lineup optimizer, which allows for tons of user customization, adjustments, locks, total ownership constraints, etc. you can learn more about our AC Plus membership here: https://analyticcity.com/pages/become-an-analytic-city-plus-member. If you already have your own, I hope this gave you some insight into strategies that might help you when using it on an everyday basis!!