World Series Game 3 was a tale of two hooks. Managers often get too much credit or blame for how their teams perform, but one of the most important decisions they can make is pulling the starting pitcher at just the right time. In Game 3, neither the Mets’ Terry Collins nor the Royals’ Ned Yost made the best decision, but one manager had his faith in his pitcher rewarded, and one suffered.Against a Royals lineup that doesn’t strike out, Mets starter Noah Syndergaard did as well as could be expected, fanning six batters. His four-seam fastball averaged 97.5 mph, higher even than the lofty baseline (97.1) he’d set in the regular season. From the first pitch of the game — which buzzed Royals leadoff hitter and magical totem Alcides Escobar — Syndergaard looked calm, purposeful, and in control.But in the top of the sixth, Syndergaard had retired 12 consecutive batters before giving up a single to Mike Moustakas and walking Salvador Perez (a rare feat) and Alex Gordon. Bases loaded, two out, Syndergaard had thrown 102 pitches — should Collins pull him? This season, Syndergaard allowed an unimpressive .860 OPS to opposing batters the third time through the order, and one lucky swing could have turned this game into a Royals win.Collins stuck with Syndergaard. It “worked” — the Royals’ Alex Rios grounded out to end the inning, and the Mets’ bullpen took care of the rest. After the game, Collins said, “I just thought that was a situation where, listen, we needed that third out and I thought he was the guy to do it.”The other guy who could’ve done it was Bartolo Colon. Colon, the savvy starter turned postseason reliever, gave up only a .698 OPS his first time through the order, much better than a tiring Syndergaard was likely to do. What’s more, that stat comes primarily from Colon’s experience as a starting pitcher, so he was likely to be even better in that situation coming from the bullpen.Meanwhile, Royals starter Yordano Ventura’s fastball wasn’t at its best, averaging 93.7 mph,1I’m using Pitchf/x data from MLB, scraped with John Choiniere’s script. almost three ticks lower than his average regular season velocity of 96.4. The loss of velocity showed; Ventura gave up seven hits in only 3 1/3rd innings of work and recorded only one strikeout.With the benefit of hindsight, it’s easy to see that pulling Ventura earlier would have been the right move. At the time, Yost may have reasoned that Ventura didn’t have much of a dropoff in the second time through the batting order, allowing a .661 OPS against opposing batters this year with a 2.5 strikeout-to-walk ratio. But Ventura was clearly having trouble, and the alternative was a Royals bullpen that was second-best in MLB at preventing runs. Even a relatively mediocre option from the pen like Luke Hochevar posted a better strikeout-to-walk ratio than Ventura in the second run through the order, and Yost could have drawn on him or converted starter Danny Duffy to throw a couple of innings each.Instead of pulling Ventura, Yost watched him give up a second homer, to Curtis Granderson in the third, before allowing three straight hits in the fourth inning. That’s three runs overall — and those three proved to be decisive, allowing the Mets to establish a lead they’d carry to the end of the game. Yost made a couple of other puzzling decisions, most obviously calling for Franklin Morales in the sixth. Morales, a back-of-the-bullpen player, gave up a few more runs and forced Yost to call for one of his vaunted relief aces (Kelvin Herrera).So we had two managers with similar decisions, but only one who suffered the consequences. That’s postseason baseball, where decisions are often judged in retrospect. Neither Collins nor Yost made the best choice in the moment, although Yost’s decision to stick with a starter whose stuff was measurably lacking is more egregious. Only Yost saw his starter fall apart, and the Royals lost a game as a result.