Postseason Managerial Report Card: Dave Roberts
I’m using a new format for our postseason managerial report cards this year. In the past, I went through every game from every manager, whether they played 22 games en route to winning the World Series or got swept out of the Wild Card round. To be honest, I hated writing those brief blurbs. No one is all that interested in the manager who ran out the same lineup twice, or saw his starters get trounced and used his best relievers anyway because the series is so short. This year, I’m sticking to the highlights, and grading only the managers who survived until at least their League Championship series. I already covered Stephen Vogt, Carlos Mendoza, and Aaron Boone. Today, I’m looking at Dave Roberts.
My goal is to evaluate each manager in terms of process, not results. If you bring in your best pitcher to face their best hitter with the series on the line, that’s a good decision regardless of the outcome. Try a triple steal with the bases loaded only to have the other team make four throwing errors to score three runs? I’m probably going to call that a blunder even though it worked out. Managers do plenty of other things — getting team buy-in behind closed doors for new strategies or unconventional bullpen usage is a skill I find particularly valuable — but as I have no insight into how that’s accomplished or how each manager differs, I can’t exactly assign grades for it.
I’m also purposefully avoiding vague qualitative concerns like “trusting your veterans because they’ve been there before.” Playoff coverage lovingly focuses on clutch plays by proven performers, but Luke Weaver and Brent Honeywell were also important contributors this October. Forget trusting your veterans; the playoffs are about trusting your best players. Mookie Betts is important because he’s great, not because he already had two rings. There’s nothing inherently good about having been around a long time; when I’m evaluating decisions, “but he’s a veteran” just doesn’t enter my thought process. Let’s get to it.
Hitting: A
The way the Dodgers were built this year, their normal pinch-hitting handedness dance was out of the question. In fact, this roster was pretty close to set in stone to begin with, and injuries to Freddie Freeman and Miguel Rojas made it more so. The Dodgers only pinch-hit twice in their entire playoff run. That wasn’t some sign of poor management; there just weren’t many good options. Give your catcher a day off? Not when Will Smith is your starter and Austin Barnes is your backup. Platoon your outfield for an extra lefty bat? Not when that bat is Kevin Kiermaier.
Luckily for Roberts, his roster was flexible enough to render Rojas’ injury inconsequential. Tommy Edman shifted from center to short, a wide assortment of utility players handled center, and the machine kept rolling. Meanwhile, Freeman’s injury meant defensive replacements and days off, but Max Muncy shifted over from third base to handle that particular necessity. This one was a triumph of roster construction more so than managing; building your team with Edman, Enrique Hernández, and Chris Taylor goes a long way towards immunizing yourself from injury-related defensive problems.
One of the two pinch-hitting opportunities was seemingly a rest decision in disguise: with two outs in the top of the seventh, trailing by four runs in Game 5 of the NLCS, Roberts pinch hit for Smith with Gavin Lux, giving Smith two innings of defensive rest. The other one was just box-checking, a ninth-inning substitution (Taylor for Lux against a lefty) in Game 4 of the World Series, with the Dodgers trailing by seven.
In being so stingy with his bench bats, Roberts avoided something that has hurt the Dodgers in past postseasons: over-platoon-itis. How many times did the Dodgers start a pile of lefties against a righty starter, hammer their right-handed platoon partners as soon as the opposition brought in a lefty specialist, and then end up with a group of noodle-bat righty bench players facing good righty relief pitching in the late innings? The days of Lux, Cody Bellinger, James Outman, Joc Pederson, and the crew filling out a large swath of the lineup are over. Now the lefty bats in the LA lineup are Freeman and Shohei Ohtani, and they’re not getting subbed out.
This approach led to the occasional weak matchup – Lux against a lefty or someone from the Hernández/Edman/Taylor/Andy Pages set against a righty – but Roberts mostly just stuck with his better players and lived with the results. I want to credit him for that, even if it seems obvious. Not every manager would be willing to sit on his hands when there was something to do, even if that something would likely end up counterproductive.
I’ll give Roberts credit, too, for making smart defensive substitutions. His roster allowed him to, of course, but he took full advantage. He consistently put good defensive teams onto the field when he had the lead. That’s not hard, but you still have to do it, and he passed with flying colors.
One minor quibble: I thought Roberts called for too many bunts. In Game 1 of the NLCS, the Dodgers sacrificed twice, neither of which I loved. In Game 1 of the World Series, Hernández bunted with runners on first and second and no one out, bottom seven, trailing by a run. That’s actually one of the best spots to bunt, especially with sinkerballer Clay Holmes pitching, so I actually liked this one. Finally, in Game 3, Edman tried a sneak attack squeeze bunt with runners on the corners and no one out. His bunt was quite poor, and the pitcher was able to shovel home for a close out at the plate (I still think Lux might have been safe, and I’ve seen the replay many times). I’m not sure I would have called that, but if I were ever going to, it would be with bunting wizard Edman in the box, with a bad matchup for him on the mound. Maybe a small deduction here, but none of these decisions bother me much at all.
It might sound like I don’t have much to say on the hitting side. That’s basically right, and that’s a good thing. Near-perfect marks from me, with a lot of that going to roster construction. It helped that the Dodgers scored so many runs that there were fewer high-leverage decisions to make, but that’s just how this team was built. All you can do is manage what you have, and Roberts did just that.
Pitching: A
I’m going to do something slightly different in this section, because I don’t know how many times I can write “the Dodgers were up 17 runs in the third inning so Roberts got to use whoever he wanted.” Instead, I’m going to give you some general archetypes of decisions that faced the team throughout their run, then dive into the specifics for each one.
The Dodgers didn’t play a lot of close games, but their bullpen management in those was on point. Take Game 1 of the NLDS against the Padres, for example. Yoshinobu Yamamoto got roughed up early, as did Dylan Cease. Yamamoto left after three innings, to be replaced by a succession of high-leverage relievers. Ryan Brasier provided a bit of bulk, Alex Vesia got the lefty chunk of San Diego’s order, and then the three-headed closer committee of Evan Phillips, Michael Kopech, and Blake Treinen recorded the final 10 outs of a three-run game.
That pattern repeated itself in Game 5 of the same series, with the top four relievers (the three righty “closers,” plus Vesia) combining for four innings of scoreless work. A modified version, with Brusdar Graterol replacing Phillips thanks to the former’s recovery and the latter’s injury, pitched the final four innings of Game 1 of the World Series (one earned run). Kopech, Treinen, and Vesia did the job in Game 2. Graterol, Vesia, and Kopech were instrumental in Game 3.
This plan isn’t complicated: find your biggest problem and throw your best relievers at it. When the Dodgers led by a small amount, Roberts consistently used a succession of high-leverage arms, and he mixed and matched those relievers enough that the same hitter never got a long run of looks at the same pitcher.
The biggest red flag I can find in Roberts’ close game management was leaving Jack Flaherty in for the sixth inning in Game 1 of the World Series. He got the top of New York’s order for a third time and surrendered a two-run homer to Giancarlo Stanton. In a postseason where Roberts generally didn’t lean on his starters – and they generally didn’t give him length when he tried to – I wasn’t in love with letting Flaherty pitch there. But the counterpoint is obvious: Flaherty was dealing, and Roberts didn’t yet know what he’d get out of Graterol in his first game action. I would have used Kopech there, personally, but I think it’s a close call either way.
Another pillar of Roberts’ approach was planned bullpen games. With only three healthy starters available, there was no getting around it. The first one came in Game 4 of the NLDS, with the team’s collective back against the wall. Roberts used his two lefties (Vesia and Anthony Banda) to attack a pocket of left-handed hitters at the bottom of the Padres lineup. He split up his top relievers and used them every time Fernando Tatis Jr. batted. By the time he ran out of elite bullpen arms, it was 8-0 in the ninth and Landon Knack could take things home.
That’s what the Dodgers bullpen game plan looked like when the team was at full strength and the series was on the line. The next time that spot in the rotation came up was Game 2 of the NLCS, and this time Roberts was down a top reliever (Vesia). With more margin for error in the series and less ability to use premium arms for the entire game, Roberts made a change that I think is smart. He inserted Knack into the game in the second inning and tried to figure out what he had.
Think of it this way: There was pretty much no way to throw a bullpen game without Knack taking an inning of work. The math just doesn’t add up. Even if the top three guys covered four innings between them; even if Brasier, Banda, and Daniel Hudson chipped in an inning each; even if all of that came off without a hitch, there were still two innings left. Knack was getting at least one of those and maybe two. He was always going to be the weakest link in the chain. Why not find out whether he was going to hold or break early on, and then save your best relievers if he faltered?
Knack got absolutely shelled, as it happens. Three of the first four batters he faced reached. Mark Vientos socked a grand slam two batters later. And so Roberts pivoted from a chain-of-closers bullpen game to a mop-up-the-loss version. Knack got another inning. Honeywell covered three of his own. Edgardo Henriquez pitched the eighth and ninth to close out the loss. By testing the most likely failure point early in the game, Roberts was able to better conserve his scarce resources.
The Dodgers used a near-identical plan in Game 6. Ben Casparius, another long reliever, pitched the second inning, entering after Kopech pitched a stressful first. Unlike Knack, Casparius put up a zero. Then the lefty part of the Mets lineup batted, and Roberts called on Banda. In the meantime, Los Angeles’ lineup got to work. When Banda took the mound again in the top of the fourth inning, he did so with a 6-1 lead. That was a green light for Roberts to deploy the big guns. The next time the top of the order came up, Phillips faced them. The time after that, it was Treinen’s turn. Medium-leverage relievers like Brasier and Hudson filled in the gaps. Roberts had a clear plan: let Casparius cover an inning, because he had to mathematically, and then reassess, with a strong preference for using his best relievers if they had a lead or were trailing close.
The last bullpen game of the Dodgers’ playoff run was a controversial one. Casparius drew the start in Game 4 of the World Series, and he managed two strong innings. Roberts had to balance one key factor, though. For the first time all October, he’d been riding his top relievers hard. Three straight games had been close late, and he emptied out the bullpen in every one. The previous game, played the night before, had featured six relievers. Game 4 was undoubtedly going to feature many relievers. There was another game scheduled for the next day. The Dodgers led the series 3-0.
Roberts looked at all of that and decided that if his team was winning this game, they were going to do it with Casparius getting multiple innings, Hudson going at least one and probably more, and an appearance from Honeywell or Knack mixed in there too. That plan lasted all of six batters into Hudson’s appearance, which is how long it took before Anthony Volpe launched a grand slam off of the right-hander.
I don’t think there was much Roberts could have done there. Given the shape his bullpen was in, pressing hard to try to keep this game close in the early innings just didn’t make much sense to me. After the Volpe grand slam, the writing seemed to be on the wall, so Roberts followed with Knack and just let him go. You can understand the reasoning here. Why use your best relievers when you’re worried about reliever overuse, trailing by three, and comfortably ahead in the series? But then the offense made things complicated by cutting the deficit to 5-4.
Roberts didn’t deviate from his low-leverage plan. Knack covered a whopping four innings. Honeywell got the eighth inning and ended the tense part of the contest by giving up five runs. The Dodgers never scored again, for what it’s worth, as the Yankees deployed their best bullpen arms to claim their first win of the series.
Using Knack when down by three runs? That’s just logic. But using Knack in the fifth, sixth, and seventh innings, with a single run deficit? That’s something else entirely. You can imagine Roberts going for broke and unloading his best arms at that point, heavy workload and all. You play to win the game, and all that. Who could have known that the Dodger offense, which had been mashing people all season, wouldn’t break through? Effectively conceding a World Series game halfway through is a bold move in any situation.
I’m here to tell you that I would have done the same thing. The Dodgers weren’t particularly likely to win the game when they finished the top of the fifth down by a run. It would have cost them dearly the following day to use their single-game win-maximizing strategy of pushing the elite relievers as far as they could go. Do it and miss, and you’d end up with a hamstrung unit for Game 5 the next day.
To be clear, a miss was always the most likely outcome, even when the deficit was 5-4. The Yankee bullpen is good! They threw four innings in this game when it was close, excluding the top of the ninth in an 11-4 game, and allowed a single hit. They struck out seven. No matter how good your offense is, it’s tough to score on the scary parts of the Yankee ‘pen. Not only that, but the Yankees are allowed to score more runs too. Roberts made a risky decision to save resources for the next day, one that could have looked bad if it backfired. I believe it was the correct choice.
Okay, so we’ve got aggressive bullpen management when the Dodgers were trying to close out leads or stay alive in the Padres series. We’ve got a flexible bullpen game approach that prioritizes using the weak links early and adjusting on the fly from there. That leads naturally to the closing game of the playoffs, Game 5, when all the effort Roberts spent preserving his bullpen paid off.
Flaherty started and just plain didn’t have it. Two of the first four batters he faced homered. Two of the first three batters of the second inning reached base. Now the bullpen situation was completely different: everyone was rested, and the next day was a scheduled travel day. Roberts didn’t hesitate; he pulled Flaherty after only 35 pitches and nine batters faced.
That made the game another de facto bullpen effort. Roberts managed it aggressively. He attacked the top of the Yankees lineup with his best relievers. When Graterol got into a jam in the bottom of the sixth, Roberts didn’t go to a “sixth-inning guy.” He called in Treinen, his best option, and told Treinen to pitch as long as he could. That worked out to seven gutsy outs over 42 pitches, the most pitches Treinen has thrown in a single outing in his entire tenure with the Dodgers. Fresh off of three straight days of not pitching, and with another off day coming up, Treinen was free to let it eat, and Roberts let him go.
Partially, Roberts did that because he saw Treinen was effective and leaned into it. Mostly, though, he did it to make the numbers add up. The Dodgers had used six pitchers in the first 5.2 innings of the game. After Treinen, their bullpen was thin: Casparius (43 pitches the day before), Knack (56), Honeywell (50), and Hudson (20 the day before, and 22 the day before that).
Maybe Roberts could have squeezed an extra out or two out of Kopech and Vesia, though I’m skeptical. Regardless, he had to cover the last three outs of this game, because the offense had rallied back to take a 7-6 lead while the bullpen had been dealing. From Flaherty’s unceremonious exit in the second through Treinen’s heroic stint in the sixth, seventh, and eighth, Dodgers relievers pitched 6.2 innings and gave up two runs. Now someone had to get the ninth inning.
As it happens, that someone was Walker Buehler. I have to say, I expected to hate this decision, but the longer I thought about it, the more I liked it. Buehler was in line to start Game 7 on regular rest, and he’d thrown an abbreviated outing in Game 3. Using him as a reliever here was borrowing from the team’s chances of winning Game 7, but it was doing so in a moment of huge leverage. Up by one run, three outs to get, World Series as the prize if you can do it. This is when you want to sacrifice the future for the present.
Buehler comfortably accomplished his task, striking out two Yankees in a perfect inning. But that was only possible because the offense and bullpen saved Flaherty’s bacon. When your starter gives up as many runs as he records outs, you generally lose. Only Roberts’ previous stinginess with relievers in Game 4, and his aggressive use of them in Game 5, prevented that fate.
A quick word on blowouts: the Dodgers played in a ton of them this October, in both directions. For my money, Roberts did an excellent job managing them. It’s not so much that managing a blowout is hard, but I wrote these columns for three other managers this year, and few of them could resist running their closer out there with a five-run lead every so often. That’s just bad process.
If you’re up, you let your starters run, use your worst relievers, and stand ready to bring in the heavy hitters if something goes wrong. If you’re down, you use all the lowest-leverage options you have. It’s simple, and Roberts did it right. The Los Angeles offense created a ton of blowouts this postseason, and the shaky starting pitching allowed its fair share in the opposite direction. I’m not sure how the bullpen would have held up if they’d been involved in a long string of tight contests, but that was never particularly likely given the rest of the roster.
Here’s how I see it overall: Roberts had a few guiding principles that he used to choose his pitching matchups. He hewed to them closely – so closely, in fact, that the biggest question about his decision-making is whether he should have deviated more from the plan in Game 4 of the World Series. I’m sure a lot of people will dislike his decision to care more about the future than the present in that instance. If it had backfired on him, if the Yankees hadn’t choked away Game 5 with horrendous defense and rallied to win the championship, we’d be talking about this decision in a very different light.
On the other hand, a lot of things could have gone differently. If Flaherty had made a better start in Game 5, who knows what would have happened? If Roberts had gone to his best relievers in Game 4, only for the offense to come up short, how would the important innings have gone in Game 5? If the series went seven, how would Buehler have fared? We’ll never know the answer to any of these questions.
Weighing that all together, I would have handled that key choice the same way Roberts did. I don’t find a lot to fault in the rest of his decisions. I don’t do this lightly, and I admit that it’s in the eye of the beholder, but what can I say? Roberts understood his team construction exceptionally well. He managed to maximize its strengths and hide its weaknesses. He dealt with a startling lack of innings from his starters, against a tough slate of opponents. He deployed a terrifyingly effective bullpen in every game that he led in the late innings. Sorry, noted Roberts distruster Craig Goldstein: I gave him an A.
Ben is a writer at FanGraphs. He can be found on Twitter @_Ben_Clemens.
Leaving in Flaherty could have easily cost the Dodgers Game 5. They still had Yamamoto in Game 6 so they might have been ok. But a Game 7 Buehler start would have been anxiety inducing
Yeah, I’m not sure if it was concern about taxing the bullpen (or over-overexposure), but it did not seem prudent to me, either. Otoh, Roberts has made better calls and gotten unlucky in the past, so maybe this really is just a regression to the mean for his playoff outcomes
A Yamamoto start in Game 6 might have been seen by more viewers than any other game in the history of Japan.