Skip to content
← Learn

DRBL numbers

What is DRBL?

DRBL reads public play-by-play and estimates how much each player helped or hurt his team's expected scoring, compared with a role-matched baseline. Two numbers carry almost all of it.

Two main numbers

DRBL/100How good?Impact per 100 possessions. Positive is above the baseline, near zero is about even, negative is below. This is the main ranking number.WAR1How much?Season value in wins. Minutes matter, so a starter usually beats an equally good bench player. This is not traditional WAR.

Offense and Defense on player pages are DRBL-O and DRBL-D, one for each side of the ball. They add context and are not a second ranking. A blank means DRBL isn't published for that season, not that the player was average.

Go deeper

Offense and defense

How it works

Limits and coverage

Diagnostics

These appear only in season compare. They are non-additive: DRBL-P, DRBL-LN, and DRBL-B do not sum to DRBL/100.

R1 Points and the WAR1 conversion

R1 Points are the same season credit in points instead of wins. We keep them for accounting checks; the site shows WAR1.

WAR1 = R1 Points / 37.49

The divisor is a fixed number of points per win (shown rounded), so ranking players by R1 Points or by WAR1 gives the same order.

WAR1 is not traditional WAR. The name is intended as Wins Above R1, and R1 is a role-matched baseline, not classic replacement level.

DRBL/100 itself is the raw rate pulled toward zero by an empirical-Bayes prior, so small samples can't dominate:

validatedDRBL100 = EB1600(rawAbilityRate) toward 0 · k = 1600

See it on the board

Sort by DRBL/100Sort by WAR1