One person, thirty teams, no advertising.

Why it exists

Most places that publish basketball numbers do one of two things. Either they give you the box score, which is accurate and says very little, or they give you a single rating that claims to settle an argument and will not show its work. This site tries for the space between: measures that are worth knowing, each with the context that makes it mean something, and an explanation of what it does not capture.

That is why a shooting percentage here is set against what the league makes from the same part of the floor, why a lineup rating is pulled toward the team average when it comes from ninety possessions, and why the schedule adjustment is described as a nudge rather than a rating system. A number without its baseline is a number that can be quoted into meaning almost anything.

Who builds it

Adriel Conde. The data pipeline, the metrics, the site, and the writing are all mine, which means the mistakes are too. If a number here looks wrong it probably is, and I would rather hear about it than not: my profile is at github.com/adrielconde9.

How to read it

Every metric on the site is defined in the glossary, with the formula wherever a formula is the clearest way to say it. How the data is collected, how often it refreshes, and what each derived number assumes are all set out on the methodology page.

Independence

BoxScore Lab is not affiliated with the NBA, with any team, or with any betting operator, and nothing here is offered as betting advice. Statistics come from NBA Stats, the league’s own public data. Nobody pays for placement, because there is no placement to pay for.