Throughout the World Series, we'll be providing two recaps of each game, one with a focus on the winner and the other devoting a longer look to the loser. This is the Cardinals entry for Game Four. The Red Sox edition is here.

I’m starting to sound like a broken record. “If [Mike] Matheny has an obvious flaw, it’s his slow hook with his postseason starters,” I wrote on Saturday night. “And in Game Three, it came back to bite him again.”

Maybe I should make a macro; with this series still far from over, it might save me (and other internet analysts) some time. Matheny made the same mistake in Game Four, and this time the Cardinals couldn’t recover.

Look: Lance Lynn is a pretty good pitcher. As a starter, he’s struck out a batter per inning, with roughly a league-average park-adjusted ERA. He throws fairly hard and has a big, durable build, with no recent history of arm injuries. As back-of-the-rotation guys go, that’s about the best-case scenario. There’s been some talk that the Cardinals could try to trade Lynn this winter, perhaps for a shortstop, and to make room in the rotation for some younger, higher-ceiling starters who are further away from arbitration. If they do, they’ll have no trouble finding a taker.

But Lynn has limitations, some of which are specific to him, and some of which apply to all starters. Be mindful of those flaws, and you'll minimize their effects. Pretend they don't exist, and you'll make them more costly.

Lynn threw 79 pitches through the first five innings of Game Four, striking out five Boston batters and allowing two hits, two walks, and one run. In the regular reason, when teams go a week (or more) between off days, every manager in the majors would let the 26-year-old start the sixth in a similar scenario, hoping to coax an extra inning or two out of him and save someone else’s arm.

But last night was the World Series, and the rules weren’t the same. The Cardinals had four days off between NLCS Game Six and World Series Game One, another off on Friday, and one more scheduled for Tuesday, with Adam Wainwright and Michael Wacha slated to start the next two games. Joe Kelly had given them 5 1/3 innings in Game Three, and although their five most reliable relievers had appeared after him, only Trevor Rosenthal had faced more than four batters. With the winter no more than four days away, there was no need to eke out another few outs from Lynn to keep the bullpen fresh. This was the sort of situation a manager spends the first six months of the season keeping the bullpen fresh for.

The only reason for Matheny to leave Lynn in for the sixth, then, was if he believed that the right-hander would give him a better chance to get the next few outs than one of several relievers. And it’s tough to make that case.

For one thing, the lineup was about to turn over again. Beginning with leadoff hitter Jacoby Ellsbury, Lynn would be facing batters who’d seen him twice already, and we know what happens when pitchers take a third trip through the lineup: contact rates rise, and balls get hit harder. (Historically, this hasn’t been true for Lynn, who’s actually limited batters he’s seen twice in the same game to a .616 OPS. But the sample is small, and it’s mostly based on a low BABIP; Lynn’s K/BB ratio has fallen from 3.50 in his first trip through the lineup to 2.51 in his second and 2.05 in his third.)

On top of that, the first and fourth batters due up were left-handed hitters, and the second (Daniel Nava) was a switch-hitter who’s much more effective from the left side. Lynn is a right-handed pitcher with lousy results against left-handed hitters and a profile that supports sizeable splits. Matheny could have gone to Kevin Siegrist, a lefty who’s effective against opposite-handed hitters. Instead, he stuck with his starter.

The frustrating thing about managerial mistakes—particularly those involving pitching changes (or non-changes)—is that they often end up looking deceptively like the right move in retrospect. After all, even a suboptimal matchup is usually going to go the pitcher’s way. And sure enough, despite the deck being stacked against his starter, Matheny almost got away with leaving Lynn in. The righty earned two quick outs, getting Ellsbury to pop up and Nava to ground out.

But then Dustin Pedroia singled, bringing up David Ortiz. And here Matheny made another strange decision: just as he’d stuck with Carlos Martinez against Ortiz over a warmed-up Randy Choate in Game Two, he went with Lynn against Ortiz over a warmed-up Choate last night. Lynn threw four semi-intentional balls to Ortiz, later telling reporters, “I’m not one to be dumb. I’m not going to let that guy beat me in that situation.” After that walk, with two outs and two on, Matheny finally made a move, replacing Lynn with Seth Maness. And Maness left a 2-2, 90-mph sinker up to Jonny Gomes, who drove it over the left-field fence for a game-winning three-run homer.

Matheny faced some criticism over the sequence after the game, but not for the right reasons. Here’s an excerpt from the game story at

[Lynn would] rebound to retire the first two batters of the sixth before Dustin Pedroia dropped a single into left-center. That brought Matheny to the first of two key decisions he would have to make that inning. He had lefty specialist Randy Choate available in the bullpen, though the numbers did not make that the obvious move.

In 10 career at-bats against Choate, Ortiz had four hits. Ortiz had also singled off Choate the night before. Matheny did not like the matchup.

"He was ready," Matheny said. "We just weren't going there."

The first key decision Matheny had to make didn’t come with Ortiz up. It came before the inning, when he left Lynn in. Leaving Lynn in longer, with Ortiz up, just compounded the original error. Once Matheny decided to pitch to Ortiz—as he should have, since walking him meant advancing the tying run to second—the numbers made going to Choate the obvious move, despite what the game story says. But to come to that conclusion, Matheny would have had to have been looking at more meaningful stats: not the tiny, 4-for-10 sample of head-to-head matchups between the two, but the massively more significant career performance of Ortiz and Choate versus left-handed opponents, which suggests that Choate vs. Ortiz is a matchup a manager should be happy to have. Matheny fell for the fallacy that Ortiz couldn’t be pitched to, both because the Boston first baseman had success in this series and because he’d hit a few singles off Choate. For a manager, a short memory isn’t an asset. Neither is a selective one.

Matheny is taking some flak for going to Maness after Ortiz reached base, because Maness was the one who gave up the big blow. But bringing in Maness was the best move Matheny made in the inning. The groundballer had given up only two homers to right-handed hitters this season, and Gomes had hit only five against right-handed pitchers. That matchup was unlikely to lead to disaster, but just as the wrong managerial move often ends in success, the right one sometimes backfires.

There’s a comment on that game story by a reader named “ThankYouCards.” “89 pitches, 1 earned run, Lynn was doing his job,” it says. “Walking Ortiz is hardly a reason to hook.”

In other words, if Cardinals fans (and Lance Lynn) are angry about anything, it’s that Matheny didn’t leave Lynn in longer.

The Red Sox might still have scored in the sixth had Matheny done something differently, or they might have won some other way. We can’t say the Cardinals failed to win because of Matheny, but we can say he made it more likely that they’d lose. In-game tactics are about paying more attention to process than results. They’re about knowing what the research says about how starters pitch when they go deep into games, even when they’ve pitched well up to a point. And they’re about being skeptical when it comes to small samples. Maness shouldn’t have left that sinker up, and Kolten Wong shouldn’t have slipped after taking a too-large lead. Rookies—and non-rookies, for that matter—make mistakes. But only Matheny’s mistakes were premeditated.