The Mets Bought Two Closers. Both Are Broken.

Begin with the line: Kimbrel pitched one inning. He allowed three hits, one walk, and four earned runs. Two of his pitches were strikes that found bats, the third was a slider Jake McCarthy — the Rockies’ right fielder, who entered the game hitting .241 — deposited beyond the right-field fence with the bases loaded. This was the eighth inning of a 2–2 game. The Rockies, who had been shut out by Senga and the Mets bullpen for seven, had been three outs from extra innings. They are now ten games over .500 against the Mets at Citi Field this year, which would be a remarkable sentence to write about any team but is, this week, a sentence about the Colorado Rockies, who are 15–23 on the season and were not, last week, expected to be the team that wins the season series with anyone.

This is the second issue this newsletter has written about a Mets bullpen pitcher giving up runs in the late innings. The companion piece, on the collapse of Devin Williams — the closer the Mets signed in December for three years and fifty-one million dollars, specifically and exclusively to fix this problem — ran a few issues back. The two pieces, taken together, are not coincidence. They are a pattern.

The Veteran-Closer Stack and the Reasons It Fails

For at least the last fifteen years, contending teams in baseball have run the same playbook to fix late-inning problems. They identify a former All-Star reliever in the open market. They sign him to a contract that compensates for his most recent good year, not his most recent bad one. They install him at the back of the bullpen and pray. The Mets, this winter, did this twice. They signed Devin Williams (age 31, 2024 Brewers All-Star, 2025 Yankees disappointment) to be the closer, at $51 million across three years. They retained Kimbrel (age 38, nine-time All-Star, currently older than every Mets reliever except the imaginary one in your head from the Bobby Valentine era) as a setup and emergency option. Both of them have, as of this morning, ERAs above the pitcher’s ages they replaced.

The mathematical problem with this strategy is that relief pitchers, as a class, decay sharply and unpredictably. Year-over-year correlation in reliever ERA is one of the lowest in baseball, somewhere in the .25–.35 range, which means that even very good relievers turn into bad ones with shocking regularity. Stacking two former All-Star closers does not increase the probability that you will have one good closer; it increases the probability that you will have two relievers regressing to their median, which, statistically, is exactly the spot the Mets are in right now.

Senga Pitched Seven, By the Way

It is worth noting that the starting rotation, the part of the team that does not get the headlines on a Wednesday night in May, did its job. Kodai Senga went seven, allowed two runs, struck out eight. The team led 2–1 going to the bottom of the seventh. The bullpen got an out from one reliever, allowed two from the next, and then handed the eighth to Kimbrel. None of this is Senga’s fault, and none of it is the offense’s fault for getting only two runs against a Rockies starter who pitched well. The team played well enough through seven innings to win.

The team played the eighth inning, however, and the eighth inning is when this team pays $250 million in salary to make sure things go badly.

What This Looks Like If You Squint at It from June

The honest projection: Kimbrel, even at thirty-eight, is unlikely to keep this up. Career relievers with his profile typically post some terrible stretches and some fine ones, and the 4-ER inning is, in isolation, not predictive. Williams, similarly, has had bad stretches before and pitched out of them. There is a version of this season in which both pitchers find themselves by July, the bullpen stabilizes around a 4.00 ERA, and the Mets, somehow, climb back into the wild-card conversation.

The less-honest projection: a $250 million team with two collapsing late-inning relievers will trade for someone at the deadline, sign a free agent next winter, and try this same strategy again, because there is no other obvious strategy when you have decided to spend money on a problem that does not yield to money. Late-inning relief, as the Yankees and Cardinals and Padres and now the Mets have all demonstrated, does not respond to payroll the way starting pitching or position-player depth does. It responds to the cruel arithmetic of small samples and aging arms.

Tonight, the Mets play the Cubs, who are 26–12. Senga is not pitching. The bullpen, as currently constituted, is going to throw approximately three innings. Statistically speaking, you are about as likely to enjoy them as you were last night.