What if we Modified Net Rating a tad?

Well, I did, and the results aren’t shocking

If you’re a basketball fan, you’ve probably heard the term “net rating.” If you’re reading this, you almost certainly know what it is and are a little confused as to where I’m going with this, but I promise there will be a minor payoff. Now, on to defining what we already know. Net rating is the difference between a team’s points scored and points allowed adjusted to 100 possessions. It’s a simple measure of how good a team is and is great for gauging team quality. However, net rating does have one little flaw, it treats all opponents as equal.

The NBA refuses to go to a balanced schedule because I guess some people still care about divisions (if you know one, it’s time to put them in a home), and everyone loves how the Western Conference will see worthy playoff teams be axed in the play-in just so the Bulls, Hawks, and Pistons can slap fight their way to the eighth seed in the East. Because NBA schedules aren’t balanced, it means net rating is just a tad bit unbalanced as well, and I wanted to rectify that, so I did.

The means I used to “improve” net rating weren’t complex, but I believe it elegantly solved the issue of strength of schedule. I took every team’s offensive and defensive ratings from each game and found the difference between it and their opponents’ season-long offensive and defensive ratings. From there, I could average everything out and land on my adjusted net rating. My hope was this would give me a more accurate reading of each team’s true talent level, and it did, but it also might have been a complete waste of time.

It turns out, net rating and Basketball Reference’s Simple Rating System (SRS) do a bang-up job already (they both have a linear correlation of 0.99 to adjusted net rating). While I think my adjusted net rating is cool, it’s but one sprinkle on a sundae. However, just because the adjusted net rating doesn’t tell us anything new, that doesn’t mean it doesn’t have some interesting information.

Unsurprisingly, teams at the top and bottom in net rating generally got pulled closer to the middle through the adjustment. This makes sense as more data is incorporated into the sample and that whole regression to the mean business. While the Thunder losing a half-point of net rating and the Wizards gaining 1.7 doesn’t change their fortunes, a swing of one whole point can be massive for teams in the middle.

Let’s take the Warriors as an example. They’re 25-24 with a net rating of -0.2. However, their adjusted net rating is +0.7. For a team that is currently 11th in the West but tied with two other teams for ninth, that +0.9 could very likely be the difference between hosting a play-in game and hosting a play-in watch party.

The team that I would be most worried about are the Suns. They, like the Warriors, are also 25-24 but have a net rating of -0.9 and an adjusted net rating of -1.99. So yeah, if they can’t land Jimmy Butler, it might make a lot of sense to blow this thing up.

The adjusted offensive and defensive ratings are by how many points, on average, a team impacts an opponent from reaching their season average. A +1 adjusted offensive rating means a team scores one more point than what their opponent allows on average, and a -1 adjusted defensive rating means they hold their opponent to score one fewer point than normal.

I hope this brief article was interesting because it certainly wasn’t necessary. However, I thought it would be good to reinforce the virtue of net rating as a general analysis tool. No stat is perfect, but net rating is the bread and butter for a reason– it works.