Long before the win became the most universally derided statistic in baseball, there was New York Mets pitcher Anthony Young. Young, who died this week at the age of 51, is most often remembered for the longest losing streak by a pitcher in MLB history — dropping 27 consecutive decisions. For sabermetricians, however, Young has another significance altogether — he’s the perfect example of why pitcher wins and losses are utterly meaningless.
From May 6, 1992 to July 24, 1993, Young appeared in 77 games, starting 17 of them and finishing 37. He notched 16 saves and four holds — but he just couldn’t catch a break when he was awarded the decision. He lost low-scoring nail-biters; he lost games in relief; he lost after throwing a lot of pitches and after throwing very few. A total of 448 days went by without Young’s name next to a “W” in the box score. Young finally broke through at Shea Stadium on July 28, 1993, when the Mets pulled out a 5-4 walk-off victory in the bottom of the ninth against the Florida Marlins.
Of course, Young’s long-awaited win should have happened much sooner — he pitched reasonably well during the losing streak. Although his ERA of 4.36 was 15 percent worse than the National League average in 1992 and ’93,1 Young’s fielding independent pitching, which converts baseball’s “true outcomes” (homers, walks, strikeouts) into an ERA-like number, was 3.80 during the span of his skid, which was exactly league average.
|PITCHER||TEAM||FIP-||BABIP||LOB %||W||L||WIN %|
|Kirk McCaskill||White Sox||99||.291||65.0||16||21||43.2|
Sadly, Young was undone by a tremendous amount of bad luck on things he could barely control. For starters, the Mets were really bad while he was there, losing a total of 193 games in the 1992 and ’93 seasons. The team also lost a lot of close games in particular when he pitched: The Mets lost seven more times than they “should have” based on their runs scored and allowed (according to the Pythagorean expectation) during the games of Young’s streak. He was also plagued with a tie for the 21st-worst batting average on balls in play (BABIP) of any qualified pitcher in 1992-93 — much more a marker of the Mets’ crummy defense than anything Young did wrong.
Granted, Young also had the major leagues’ worst rate of stranding base runners in 1992-93 and its fifth-worst “clutch” score, which compares a pitcher’s actual win probability added to what we’d expect from his situation-neutral stats. But both of those stats are notoriously poor indicators of how good a pitcher is — lending credence to the notion that Young pitched well and simply got historically unlucky. And the team must have known this to some extent, because a truly bad pitcher doesn’t get a chance to factor in 27 decisions — he gets sent to the minors or jettisoned.
Given the number of pitchers who take the mound each season, somebody was going to be the recipient of that bad luck — and eventually, somebody was going to lose 27 decisions in a row. Young happened to be that guy. The Mets traded him away after the 1993 season — in no small part because of the losing streak — and elbow injuries ended up derailing the rest of his career. But we’ve learned a lot about baseball since the early 1990s. A pitcher’s W-L record is no longer of paramount concern, taking a seat behind newer statistics that try to tease out how much of those wins and losses are actually his fault. This revolution came too late for Young’s career, but he will always symbolize the flaws with traditional numbers — and why more modern stats are necessary.