The New York Mets defeated the Minnesota Twins 10-8 at Citi Field on April 23, 2026, in a game that swung dramatically in the eighth inning. The DiamondIQ model's estimate gave the Mets a 42 percent chance of winning before first pitch, and the contest followed a tense path through seven innings before a frantic two-way exchange in the eighth ultimately settled the outcome in New York's favor. The Mets built early advantages with a three-run first and a three-run second, only to see Minnesota claw back throughout the middle innings and mount a serious threat late.
The turning point of the game came in the top of the eighth, when Ryan Jeffers launched a home run off Huascar Brazobán that swung win probability 36.2 percent in Minnesota's favor, pulling the Twins to within striking distance and briefly flipping the game's momentum. That blow, combined with a four-run eighth for Minnesota overall, represented the most dangerous stretch for the Mets. However, Bo Bichette answered decisively in the bottom half, lacing a double off Anthony Banda that shifted win probability 33.5 percent back toward New York and proved to be the game's single most consequential offensive moment. Andrew Morris closed the inning for the Mets, though Ronny Mauricio's flyout cost New York 11.4 percent in win probability before the frame ended. Earlier, Brett Baty's strikeout in the second inning against Joe Ryan had added 11.4 percent to the Mets' chances by ending a Minnesota threat.
Bichette finished as the game's top performer by the DiamondIQ model's estimate, posting a cumulative WPA of plus 40.3 percent and an RE24 of plus 4.2, edging Jeffers, who contributed plus 34.6 percent WPA and plus 3.3 RE24 despite pitching for the losing side. Baty added plus 23.2 percent WPA and plus 2.4 RE24 for New York. On the mound, Tobias Myers was the most impactful arm of the night, generating plus 16.6 percent WPA for Minnesota, though it was not enough to overcome the Mets' offensive response in the decisive eighth inning.
Analysis generated by DiamondIQ's model.