Create a Custom NBA Metric | Design your own advanced metric for ranking the best players in the NBA.

Introducting The CUSTOM Metric

One of the first pieces we did for PerThirtySix was the Hustle Index, which was a tool where you could build out your own metric comprised things that don't show up on the box scores, like deflections and screen assists. We're happy with the visualizations and interactions we landed on with that piece and decided to build it out for more traditional stats. With that, we'd like to introduce you to the CUSTOM Index (Customizable User Stats from Transparently Obtained Metrics)!

I'm actually generally not a fan of "all-in-one" metrics which attempt to boil down the entirety of a player's value into a single number. They do make things simpler, but they strip away so much context and the weights of specific stats in these metrics' formulas are often arbitrary. We think our take on it has a few advantages:

Transparency - All of the stats in the CUSTOM Index are laid out with our beeswarm visualizations.
Customizability - You can tweak any of the components of the CUSTOM index and see how it updates your metric in real time.
Using the Distribution - Our formula uses the fact that stats tend to follow a normal distribution. Instead of using raw numbers, like a player averaging 10 rebounds per game, we ask "how much better than average is 10 rebounds per game?" and use that value in the CUSTOM Index.

Note

The table below shows data for players ranked by your CUSTOM Index. You can adjust the CUSTOM Index by using the sliders to assign weights to the different component metrics. In other words, let us know how important you think a stat is when you see and we'll update the CUSTOM Index.

You can select a player from the dropdown, click on them in the table, or click on any of the charts on the bottom to get a clearer picture of any player's stats. Hovering over the player on the charts will show tooltips revealing more information.

How is the CUSTOM Index calculated?

The CUSTOM Index for a player comes from calculating the z-score for each component metric, multiplying them by the selected weights, summing them all together, and then normalizing that number such that the highest score is 100. Note that we are not treating this as a definitive methodology or statistic, but instead something that's fun to play around with!

Create Your Customizable Box Score Stat

Select a season:

We're just getting started.

Subscribe for more thoughtful, data-driven explorations.