The Los Angeles Angels walked off the Seattle Mariners 8-7 in eleven innings on April 5, 2026, at Angel Stadium, with the DiamondIQ model's estimate moving from a 54 percent pre-game home win probability all the way to 100 percent by the game's final out. The Angels overcame a three-run Seattle burst in the fifth inning, when Cole Young launched a home run off Brent Suter that swung win probability 32.0 points in the Mariners' favor and briefly tilted the game toward the visitors. Seattle kept the pressure on into the ninth, where Julio Rodríguez delivered a single off Sam Bachman worth a 31.5-point swing, knotting or extending the Mariners' threat and forcing extra innings. The Angels, however, answered each Seattle surge, and the contest stretched to eleven frames with both clubs unable to pull away through the middle innings, each side going scoreless from the sixth through the eighth.
The decisive sequence came in extras, where Los Angeles gradually suffocated Seattle's chances. In the top of the tenth, J.P. Crawford grounded into a double play off Shaun Anderson, a swing of 16.6 points in the Angels' favor, and Josh Naylor followed with a flyout worth another 16.4 points, stranding Seattle and preserving Los Angeles's position. Bryce Teodosio then put down a sacrifice bunt in the bottom of the tenth off Matt Brash, advancing the winning run and swinging win probability 23.7 points toward the Angels.
Among the standout performers by the DiamondIQ model's metrics, Adam Frazier led all batters with a WPA of plus 31.7 percent and a RE24 of plus 1.4, while Teodosio contributed plus 23.7 percent WPA on his pivotal bunt. Rodríguez posted plus 23.4 percent WPA despite finishing with a RE24 of minus 0.9, reflecting how his ninth-inning single generated leverage that ultimately went unrealized. On the mound, Ryan Zeferjahn was the Angels' most impactful arm, adding plus 14.4 percent WPA, with Cole Wilcox contributing plus 7.7 percent as Los Angeles's bullpen closed out the marathon victory.
Analysis generated by DiamondIQ's model.