And Now, the Worst Team Defenses

It’s tough not to pick on the Cardinals these days. Last season, they won 93 games and took the NL Central title with a team that combined strong offense, exceptional defense — long a St. Louis tradition — and good pitching; it was their 15th straight season above .500 and fourth in a row reaching the postseason. This year, however, they’ve spent time as the NL’s worst team, and while they’re now merely the third-worst, at 33-46 they’re going nowhere and impressing nobody.
A big and perhaps undersold part of the Cardinals’ problem is the collapse of their vaunted defense, which has often featured five players — first baseman Paul Goldschmidt, third baseman Nolan Arenado, outfielder Tyler O’Neill, and multiposition regulars Brendan Donovan and Tommy Edman — who won Gold Gloves in either 2021 or ’22. Manager Oli Marmol has been tasked with shoehorning hot-hitting youngsters Nolan Gorman and Jordan Walker into the lineup at comparatively unfamiliar positions, as both are blocked by Arenado at third base, their primary position in the minors, and between injuries and offensive issues, lately Edman has been patrolling center field instead of the middle infield. Backing a pitching staff that doesn’t miss enough bats — their 21.1% strikeout rate is the majors’ fifth-worst — it’s all collapsed into an unhappy mess.
Given that context it’s less than surprising that the Cardinals show up as one of the majors’ worst defensive teams using the methodology I rolled out on Thursday to illustrate the best. For that exercise, I sought to find a consensus from among several major defensive metrics, namely Defensive Runs Saved, Ultimate Zone Rating, and Statcast’s Runs Prevented (which I’m calling Runs Above Average because their site and ours use the abbreviation RAA) as well as our catcher framing metric (hereafter abbreviated as FRM, as on our stat pages), and Statcast’s catching metrics for framing, blocking, and throwing (which I’ve combine into the abbreviation CRAA). Each of those has different methodologies, and they produce varying spreads in runs from top to bottom that owe something to what they don’t measure as well as how much regression is built into their systems. Pitchers don’t have UZRs or RAAs, for example, and the catching numbers are set off in their own categories rather than included in UZR and RAA. I’ve accounted for the varying spreads, which range from 86 runs in DRS (from 42 to -44) to 25.6 runs in FRM (from 13.8 to -11.8), by using standard deviation scores (z-scores), which measure how many standard deviations each team is from the league average in each category. Read the rest of this entry »