Kevin Cash Changed Bullpen Plans
The Rays made it to the World Series for a lot of reasons, but one of them is indisputably their bullpen, which has given Kevin Cash the flexibility to pull starters whenever he wants and follow them with an unending stream of hard-to-hit relievers. Cash, in turn, has used it masterfully; he’s pushed the right button at seemingly every turn. Last night, I think that might have changed.
In the top of the fifth inning, Tyler Glasnow couldn’t find the zone. He walked the first two batters he faced, allowed two runs (on a fielder’s choice that didn’t get anyone and a single), and generally looked gassed. Cash went to his reserves and brought in Ryan Yarbrough.
That sounds like a reasonable usage choice, but it’s simply not how Yarbrough is deployed most of the time. Here are the particulars of his previous playoff appearances this year:
| Game | Inning In | Outs In | Pitches | Batters Faced |
|---|---|---|---|---|
| ALDS G4 | 2 | 2 | 65 | 21 |
| ALCS G3 | 1 | 0 | 82 | 21 |
| WS G1 | 5 | 1 | 19 | 4 |
Okay, there have only been two of them, but he’s been used as either a starter or a bulk guy in both. He throws a near-starter number of pitches and faces a small sliver of the lineup a third time through. That’s similar to his usage this regular season:
| Date | Inning In | Outs In | Pitches | Batters Faced |
|---|---|---|---|---|
| 7/25 | 1 | 0 | 69 | 21 |
| 7/30 | 1 | 0 | 87 | 25 |
| 8/5 | 1 | 0 | 82 | 26 |
| 8/10 | 1 | 0 | 77 | 20 |
| 8/15 | 1 | 0 | 51 | 13 |
| 8/21 | 1 | 0 | 97 | 27 |
| 8/28 | 1 | 0 | 79 | 23 |
| 9/8 | 1 | 0 | 70 | 17 |
| 9/15 | 2 | 1 | 86 | 22 |
| 9/20 | 1 | 0 | 100 | 28 |
| 9/26 | 1 | 2 | 35 | 12 |
Mostly starts, with a few relief appearances thrown in. It’s not unreasonable that the Rays might want to turn him into more of a reliever in the World Series, the only round of the playoffs with off days, but 19 pitches? Four batters faced? The last time he faced four or fewer batters in the regular season was July 13, 2018. He had two short appearances to only one extended stint in last year’s ALDS, but that was part of a gambit to use a true bullpen game (Diego Castillo drew a start) with Yarbrough handling two innings, then use him as a LOOGY in two other games.
Yarbrough is essentially a starter. We had him penciled in for Game 4 of the World Series, something which would be tricky now; that would be on three days’ rest, and while he only threw 19 pitches, it’s still a disruption to his routine. The Rays still could use him there, but I think that game is now more likely to be a bullpen game with perhaps two innings out of Yarbrough. Glasnow, Snell, and Morton would then each draw two starts to fill out the full complement of seven games.
That plan is reasonable, if different from the Rays’ prior usage of Yarbrough. It could be that the team had already decided they only wanted 30-40 pitches out of him in Game 4, which frees up his usage in earlier games. Though he’s effective as a starter, he’s particularly good against lefties. That’s not a skillset the Rays lack — they also have Josh Fleming, Aaron Loup, Shane McClanahan, and Ryan Sherriff on the roster — but extra lefty relievers against the Dodgers always have their uses.
If that were the only weird part of Cash’s decision to go to Yarbrough, I’d understand it. There’s something much weirder, though: the Rays were already more or less dead in the water when Yarbrough entered. They trailed 4-1 with a runner on third base and less than two out, and they had only four innings to come back from it. Per our win probability model, they had an 8.2% chance of winning the game when he took the mound.
Three runs might not sound like a lot, but teams mostly don’t make up three-run deficits. One of Tampa’s quiet idiosyncrasies is their willingness to concede games like this. Take a look at the relievers the Rays have used when they’re quite unlikely (20% or less) to win this postseason:
| Game | Pitcher | Inning | W% |
|---|---|---|---|
| ALDS G1 | John Curtiss | 9 | 16.50% |
| ALDS G1 | Shane McClanahan | 9 | 0.30% |
| ALDS G4 | Aaron Slegers | 7 | 7.70% |
| ALCS G4 | José Alvarado | 7 | 11.30% |
| ALCS G4 | Aaron Slegers | 8 | 6.10% |
| ALCS G6 | Shane McClanahan | 6 | 13.70% |
| ALCS G6 | José Alvarado | 7 | 1.70% |
| ALCS G6 | Aaron Loup | 8 | 1.10% |
| ALCS G6 | John Curtiss | 9 | 3.40% |
There’s really no way around it: these are the relievers the Rays trust least. Alvarado and Slegers didn’t make the World Series roster, McClanahan has barely pitched, and Curtiss is pretty low in the bullpen hierarchy. Have you ever wondered how the Rays run out an endless string of bullpen monsters in every close game? It’s because they never use their best guys when the game is already mostly lost.
This is a change that more teams should make. For all the bellyaching about managing for today and figuring tomorrow out tomorrow, putting good pitchers in when you’re unlikely to win even if they excel is throwing good money after bad. Nick Anderson only has so many bullets per series in his arm; Cash is great at using those bullets when they count.
Maybe Yarbrough somehow counts as one of those bad relievers, but I don’t think so. The Rays have used him to get large swaths of tough opposing lineups out this postseason. They’ve used him for multiple innings in close games with other relievers available. He’s their fourth starter, essentially, which is pretty high up the bullpen pecking order after the starter/reliever translation.
For whatever reason, Cash changed tacks last night. Maybe he considered last night a use-it-or-lose-it situation — if Yarbrough is going to throw, say, 40 pitches in Game 4, he might be unavailable in Games 2 and 3. Maybe the extra off days have widened his ranges for using his best relievers; after all, there’s a day off after tonight’s game, and Anderson, Castillo, and Pete Fairbanks are all available to back Blake Snell. Maybe he isn’t willing to concede any World Series games, even if he was okay with it in earlier rounds.
Whatever the reason, though, the Rays are trying something new. I don’t think it was a great decision — nothing Yarbrough could have done in that situation was likely to matter, and I’d rather toss Aaron Loup in there and keep Yarbrough fresh in case he needs to throw more innings in Game 4. Keep an eye on Cash’s bullpen usage tonight; it’s been incredibly valuable for the Rays this postseason, but it looks like the specifics might be changing.
Ben is a writer at FanGraphs. He can be found on Bluesky @benclemens.
What was the win probability when Yarbrough began warming up? It’s not as if you can bring in a cold pitcher.
Yeah, you saw how Floro fared when they rushed him in from the pen. Still though, the larger point of burning your best arm at the least significant moment is hard to deny…
I believe we actually chatted about this, in a roundabout way, last night. Clearly, they still thought they had a chance to win. There is just no other explanation for bringing him in.
Of course, if they were going to bring him in, it would have made a lot more sense to do it in the 4th inning. But hey, what’s at stake except a possible world series championship?
When he was warming up i knew it was over. they didnt want to expend any of the A bullpen and it was too early for curtiss. so who had a inning to burn? Yarbs who would have thrown tuesday anyway on his rest day for a workout, this was a garbage time inning from a non garbage time pitcher if that makes sense.
Yep. The old “let a starter relieve an inning on his throw day” gambit
Back in the days when pitching staffs consisted of usually 8 men, for example the 1951 Yankees only had 7 pitchers appear in 20 or more games, it was routine for the starters to throw in relief. I can’t call him my favorite player, after all he pitched most of his career for the Yankees, but Allie Reynolds was a special pitcher. In a 10 year stretch between 1943-1952 he made over 20 starts in each year and recorded saves in 9 of the years. He was definitely the “Superchief”.
That is why baseball is so awesome. You are talking about pitchers from 1943-1952, and my parents weren’t even alive till many decades later, yet we are here, both reading and talking about baseball.
And sometimes what is old is new again.
I feel like the use of Yarbrough was less controversial than the overuse of Glasnow. If you felt the game was important enough to use one of your A-relievers, why not bring him in before Glasnow lost his command, threw so many pitches, and buried your team beneath a 3-run deficit? Either punt the game and use your B-relievers or be aggressive and pull your ailing starter before the sky falls in, right?
it wasn’t so much the relief pitcher sequence that stunned me as it was Cash leaving Glasnow in for 112 pitches. Glasnow was exactly sharp and mowing down the Dodger hitters, In the chat I predicted that there was 0% chance that he would make it through the 5th which was hardly clairvoyant. That is completely out of character for Cash who usually has his first pitcher out of there, even if he is cruising, well before 112 deliveries.
My thought as well.
Seems to me if you have your starter approaching 100 pitches after 2 straight walks (it was 99 after those) in a 1 run game at the time, it may be the time to make a change to pitch to Turner. Or you could ride him a bit (Glasnow did K Turner, then came the get nobody fielders choice play on Muncy’s ground ball) yet.
My thinking is either 1) bring Yarbrough in to face Muncy (103 pitches by Glas), then Smith and Bellinger (2 lefties to start), or you compete aggressively and 2) bring in Castillo (of course that may not have been an option if he wasn’t warming) in a fireman role to pitch to either Turner than onward through Muncy and Smith (2 righties) at a 2-1 score or 3) bring him in to face Smith (after 106 pitches by Glasnow) then Belli and Taylor at 3-1. Castillo hadn’t pitched since Friday. Key game moment demands a key guy.
BTW only 7 of Glasnow’s last 13 pitches were strikes. – 2 of the last 4 strikes to Muncy and Smith were put in play and drove in runs.
Yarbrough came in (after not pitching for a week) to pitch to lefty Belli, then Taylor and Keke as a righty pinch hitter for lefty Pederson. Both righties got hits to drive in 2 more runs. There likely would have been later opportunities to get Yarbrough work facing 2 lefties in 2 batters later – either Seager and Muncy or Muncy and Belli who might have been less likely to be hit for.
I would have gone with 1 of the Castillo options (or maybe Thompson who was plenty rested as well?)
They might have been planning to use him instead if having him throw his normal bullpen session between “starts”. This could be especially so since there would likely be a longer than typical gap between his “starts”.
Could it be that the world series is a different animal than a regular season game and Cash can’t afford to “punt” a game to the superior Dodgers when he only has four to give? And Glasnow likely has at most one more game to pitch this season, so stretching to 100 pitches, or Yarbrough making an early appearance in a losing situation isn’t that big of a deal.
This was my thought. The regular season is a marathon (even this year, sort of) and the post season is a sprint. You can punt a game in July to prevent a disadvantage in the next two, but that is really risky in a best of 7 series, especially when playing a better team.
“Only” an 8.2% chance to win? Somebody comes back from that almost every day during the season and most of those teams are worse than the Ray’s. I’m a big fan of managing all out and worrying about tomorrow, tomorrow in the playoffs. Everybody will get plenty of rest when it is over.
But I agree it’s a really tough call and you can definitely put yourself in a bind. The Ray’s have a pretty deep bullpen so, despite the good analysis, I’m not sure they did.
You can read more about his decision on his blog, Cash Considerations.
I think that part of the reason why Cash went to Yarbrough was because Yarbrough hadn’t pitched in a week, and Cash wanted to give another day’s rest to most of the bullpen. The last game Yarbrough pitched in was Game 3 of the ALCS, so he was on 6 days rest at that point. He threw 19 pitches and faced 4 batters. If he had gotten two quick outs, the score would have been 3-1, and continuing to use Yarbrough would be a defensible way to keep the game close.
As it was, Yarbrough allowed the score to get to 6-1 before the end of the inning. At that point, the smart play is to pull Yarbrough and allow him enough rest to come back for Game 3 or 4. It was a move that didn’t work out. Cash didn’t compound the error by leaving Yarbrough in the game.
It was a tad weird. Regardless of what the numbers say, it does feel like the Dodgers bullpen seems easy to score on. Issue is that it is also the Dodgers offense, so even down 3 you are going to have to score more than 3 runs, most likely.