DRS

Simple explanation

Picture a center fielder sprinting to catch a ball that other center fielders catch only 70% of the time. He gets +0.3 for the catch, and −0.7 if he misses it. Add that up over a season and convert it to runs, and you have the core of DRS.

Beyond range, it scores throwing (holding and throwing out runners), double plays, handling bunts, robbing homers at the wall and controlling the running game, and adds the pieces together. +5 means five fewer runs allowed than an average player at that position.

Benchmarks

−15

Very poor

−10

Poor

10

Good

15

Gold Glove caliber

Rough cutoffs for one season, comparing players at the same position (FanGraphs rules of thumb).

Detailed explanation

It's built on batted-ball location and type data recorded by hand by Baseball Info Solutions (BIS), so some scorer error comes with it. That's a different starting point from OAA and FRV, which come from Statcast tracking.

The idea matches FanGraphs' UZR, with slightly different math. DRS rates each play's difficulty on a rolling one-year basis while UZR uses several years of data, and DRS also covers pitchers and catchers. So the two sometimes disagree about the same player.

A few months of data bounce around a lot. FanGraphs says it takes one to three years before it becomes fairly reliable.