Percentiles, and What “Value+” Means

There will always be new stats and new ways to analyze players. Points per game, PER, BPM, VORP, TPA, POE, PIPM, RPM, and many more.

It’s hard enough keeping track of what each of them does. Knowing how to read the stat shouldn’t be hard. With tools like percentiles and the “Value+” stats, you should be able to understand how good a player is at something, no matter what the stat is.

What Is a Percentile?

Percentiles are simple. They show the percentage of a sample the player is at something.

If Lonzo Ball is in the 60th percentile at stat XYZ, that means he has a better than 60% of the competition in stat XYZ.

Mathematically, there is no 100th percentile. You can’t be better than yourself. So the best you’ll see for a player is the 99.9th percentile.

Percentiles make it simple to essentially rank and compare players, but they don’t capture magnitude. Players might be next to each other in rank (and thus, percentile), but be on two different tiers altogether. That’s why we at BBall Index have decided to use Value+ stats.

What Is “Value+” and How Is It Different?

Value+ numbers answer the question of magnitude. This idea was taken from baseball analytics, where stats like OPS+ are used to serve the same purpose.

The “plus” part of the stats adjust so a score of 100 is league average, and 120 is 20 percent better than the league average.

If Ian Clark’s RPM+ is 80, that means he’s 20 percent lower than league average for that given stat.

By using percentile and Value+ together, you’ll understand how a player compares to their peers from both rank and magnitude standpoints.

Tying It All Together

Here’s an example. Steph Curry was in the 96th percentile in touches with a Touches+ value of 264, meaning he had more touches than 96 percent of NBA players and had 164 percent more touches than the average player.

You can take these tools of contextualizing data and apply them elsewhere as well. But at least on BBall Index, you’ll always know how to read the data and won’t ever need to memorize what good POE or PIPM values are.

Leave a Reply

You must be logged in to post a comment.