This article was originally published on February 1, 2018.

Over the last few months, NSAG, in conjunction with Sportlytix, analyzed the DFS data for all 3,362 games during the 2016-17 NBA season. We believe our conclusions are extremely beneficial for DFS players and we will be posting a series of blog posts detailing our findings over the next few weeks. Our approach was simple: look at the data using an outside-the-box approach, and unlock the secrets that top DFS performers are using, but that are not publicly available. We would like to thank Sportlytix for all their help over the last few months collecting the data and helping us reach tangible insights.

DFS Explanation

Daily Fantasy Sports (DFS) is a growing industry globally and saw over $3 billion in entry fees in 2017. The premise of NBA DFS is simple: draft a lineup of NBA players that you believe will perform well statistically. There are a number of contest structures in DFS that can be separated into categories: cash games and tournaments. Double-ups are an example of a cash game, where players pay an entry fee, and roughly the top scoring half of all competitors win double their entry fee. Tournaments, the most popular type of contest, feature anywhere from 100-100,000 competitors and only the top 10% of performers receive payouts. The payouts are skewed towards the top positions. Below is an example of a payout structure from a $9 NBA contest with 52,287 entrants:

Sample FanDuel Prizing Payout

There are a number of theories that can be used to construct the perfect lineup. We believe we can use extensions of analysis, like application of game theory, as well as other theories to construct better lineups and ultimately perform well in contests. Over the next few blog posts, we will explain some of our findings over the past few months, and how they can be executed through different strategies.

Our Data

The data we analyzed recorded every performance from the 2016-17 NBA regular season and included basic info (player name, position, team, points, salary, etc.). A sample of the data can be seen below.

Sample data NSAG analyzed

Cleaning the Data

To eliminate outlying performances, we filtered our data to include only performances in which players played at least 5 minutes in each game. While we recognize that analyzing projected minutes played is a big part of the game, we felt that any player playing under 5 minutes would skew the data leading us away from tangible insights instead of towards them.

Pts/$K Indicator

Our most important statistic we used throughout our analysis was Pts/$K. This is calculated by dividing the total fantasy points a player recorded by their salary. This statistic is important because it allows us to compare players in terms of their efficiency. Throughout our studies, our goal was to find inefficiencies in pricing of players that had certain characteristics. The higher the Pts/$K, the more undervalued the player. The following benchmarks are good to know:

Pts/$K general ranges and descriptions

We hope you enjoy the following series of blog posts, designed to provide outside-the-box insights surrounding NBA DFS, using data from the 2016-17 season.