The Mets got better at preventing runs right when it stopped mattering

Here is the uncomfortable number at the center of the Mets' final stretch: 3.8 runs allowed per game over the last ten games, down from 5.3 over the prior twenty. A drop of a run and a half per game is not noise. That is a real, meaningful shift in how this team has defended—and it arrived approximately five months too late to save anyone's season.

The contrarian instinct here is to resist the easy eulogy and ask what that number actually means. All season long, the Mets' pitching story was volatility. Sean Manaea posted a 4.88 ERA across 36 appearances. Freddy Peralta was similarly uneven. The rotation leaked runs in April—7-19 that month, the team's worst stretch by a wide margin—and never fully recovered. The one genuine bright spot was a relief arm that put up a 0.39 ERA over 14 games, which is an extraordinary number in any context, but one reliable late-inning arm cannot paper over a staff that spent most of the year giving games away. So when the run-prevention numbers finally tighten in the closing days, the question worth asking is: what changed, and could it have changed sooner?

The honest answer is probably not. Injuries were the load-bearing wall that collapsed under this roster. Tylor Megill, Reed Garrett, and Jorge Polanco were all lost for extended stretches. Christian Scott, who had posted an 11.1 K/9 rate and looked like a legitimate rotation piece, went down on September 22. When you lose that much pitching depth, the staff that remains gets overextended, and overextended pitchers allow runs. The defensive tightening at the end may reflect nothing more than a healthy-enough group finally pitching in a stable configuration—the roster, at last, not actively on fire.

What makes the number worth preserving, though, is what it says about the floor. This team finished 73-87, a step back from 83 wins in 2025 and 86 in 2024. The offense was narrower than it needed to be—Carson Benge was the most reliable bat, Francisco Lindor hit .231 in a year that demanded more from him, and the lineup leaned heavily on a Juan Soto who was day-to-day for too much of the second half. But a pitching staff capable of holding opponents to 3.8 runs a night is not a staff that needs to be torn down. It is a staff that needs to stay on the field.

My read going into the offseason is this: the Mets should be more interested in that late-stretch run-prevention number than in the final record. Seventy-three wins is the headline, but it was built on a foundation of absences. The April collapse—twelve consecutive losses at one point—set a deficit the team spent the rest of the year clawing back from and never fully erased. That is not a talent indictment; it is a depth indictment. The talent, when present, showed up. The question for whoever is rebuilding this roster this winter is whether the defensive competence glimpsed in these final ten games represents a genuine baseline or a small-sample mirage. Given the injury context, I lean toward the former. Numbers earned while depleted tend to be understated, not inflated. The Mets were better than their record when healthy, and better at run prevention than their season suggests when the staff was intact. That is a more useful starting point than a 73-win season implies.