Is Derek Shelton Pulling Twins Relievers from a Hat?

[Greggory Masterson]

Image courtesy of © Matt Krohn-Imagn Images

The 2026 Minnesota Twins bullpen is markedly different than those of past years. Obviously, there’s a clear talent gap, but beyond that, the arms are being deployed in a seemingly inconsistent manner. Minor-league veteran Luis García threw two innings of 9th-inning mop-up to begin his Twins career (allowing two runs in two innings across two games), and the next day, he was trusted to throw the 8th inning of a 4-3 game that the Twins were winning.

In April, the Twins had back-to-back games (which they won) that required four relievers to record eight outs and five relievers to record 10. Justin Topa and Kody Funderburk both picked up saves, despite neither pitcher starting the 9th inning. To date, eight Twins have a save on their ledger, with none having more than two (Topa, García).

As stated, this pattern is a far cry from the typical usage in prior seasons. It’s much easier to set it and forget it when the backend contains Jhoan Durán, Griffin Jax, Louis Varland, and Brock Stewart. Sure, there may be some shuffling required, but whatever day-to-day usage variance existed with a more structured pen, it doesn’t hold a candle to what we’ve seen from Derek Shelton’s relief corps.

From the outside, this seems like the Twins are relying on gut instinct and hot hands to try to find a way through the late innings of every ballgame. (So up yours, analytics wonks!) However, it’s a bit naïve to believe that Shelton’s ability to white-knuckle his way through ballgames can all be chalked up to his baseball acumen and willingness to switch to another arm as soon as things start to skid.

Before every game, professional scouts and analysts prepare game plans and scouting reports to prepare players for action. They also prepare material for coaches, and Shelton has confirmed that he has long meetings about the best possible plans to get through a game with his coaching staff, just as Rocco Baldelli did.

Most teams have some hierarchy in the bullpen, all things being equal. But part of a pro scout's or analyst's role is helping managers make informed decisions within that hierarchy. One of the tools many organizations make available to managers is an in-game bullpen hierarchy. This tool might come in several forms. Many teams use a (close your ears if you hate nerds) laminated notecard listing the opposing team’s lineup. For each spot in the lineup, relievers are organized one through eight, based on the quality of the matchup with that specific hitter.

Last season, that hierarchy might not have seen much variation. The best matchup against any hitter was Durán. After that, it was Jax. If the hitter was a lefty, third might be Danny Coulombe, but otherwise it’d be Stewart and Varland, in some order. Given the caliber of arms we’re talking about, throw the best one available.

If pitchers are more evenly matched, though, that hitter-to-hitter differentiation might be more obvious. Naturally, there are platoon splits to be aware of—a guy like Taylor Rogers or Anthony Banda will get the nod against a lefty over someone like Cody Laweryson or Yoendrys Gómez. But there are also more idiosyncratic factors to consider. For instance, arm angle, pitch mix, velocity, swing trajectory and/or plate discipline can come into play.

With a bunch of pitchers who can charitably be called “fine,” chasing the matchups becomes more useful and necessary. Take Topa (who has been designated for assignment since the writing of this piece, but who knows, he may remain in the organization) and Eric Orze. Both pitchers are righties who throw their fastball in the low 90s, but they have very different profiles. Topa is a sinker-sweeper/cutter pitcher who throws sidearm and mostly operates east-west. Orze has a much higher arm angle, and over half of his pitches are splitters. When he’s not throwing a splitter, you’ll probably see a fastball, giving him a much more distinct north-south profile.

You should be able to see where I’m going here. Some hitters struggle more with north-south, and others struggle with east-west. There will be righties in an opposing lineup who should have Orze deployed against them, and others whom Topa would best match.

Tyler Duffey referenced this type of system in a Twin Cities Territory episode last month, relaying how, when he was with the Twins, relievers would be prepped before each game on which part of the opposing lineup they would likely be deployed against. This scouting seems even more vital when the bullpen has to fight and claw for every advantage possible.

So, in practice, to an outsider, that can look like Shelton pushing buttons at random. Topa may be trusted to close one day and throw the sixth the next, based on where his advantage is in the opponent’s lineup. And that quick hook for relievers may be a jumpy manager, or it might be that Cole Sands had the best matchup for the previous three batters, but Andrew Morris’s pitch mix is a better matchup for the next two, and Shelton wants to take advantage of that. Don't forget to factor recent workload in, too! There are always more layers to these decisions than meet the eye; teams work hard to keep it that way.

Of course, there will still be stuff thrown at the wall to see what sticks. They don’t know what they have in John Klein or Kody Funderburk until they’ve been tested, and Shelton can also do a little Minnesota politicking (that is, passive aggression) if he believes it will compel the front office to make a roster move he feels is overdue. All told, though, the method is probably outweighing the madness in this regard. They have a plan, even if they’re working with lesser material. One fun way to pass the next month might be to figure out what patterns we can spot in that plan, as a few pitchers settle into their roles with the club.

View full article

[Possumlad]

Not sure what the other options would be with this ragtag collection. It looks random because the results have been random (and bad). The results wouldn't look any better if you shuffled the bad pitchers in to more "defined roles." 

[mark allen]

has to be tough on the bull pen not knowing some type of order. They all half to be ready at any time, it would be nice if we could figure out who gets at least the 8th and the 9th inning. Think that would help them all be a little more prepared.

[Greggory Masterson]

5 minutes ago, mark allen said:

has to be tough on the bull pen not knowing some type of order. They all half to be ready at any time, it would be nice if we could figure out who gets at least the 8th and the 9th inning. Think that would help them all be a little more prepared.

I think guys knowing when they should expect to pitch is valuable. But I struggle to justify the idea that taking them “You will be facing the 5-6-7 hitters” is markedly different than “You will be pitching the 8th”

[Patzky]

13 minutes ago, Possumlad said:

Not sure what the other options would be with this ragtag collection. It looks random because the results have been random (and bad). The results wouldn't look any better if you shuffled the bad pitchers in to more "defined roles." 

The Dismal Science.

[gil4]

Quote

Most teams have some hierarchy in the bullpen, all things being equal. But part of a pro scout's or analyst's role is helping managers make informed decisions within that hierarchy. 

"There are no good options - just pick names out of a hat" was probably in the report.

[karcherd]

At some point, they need to determine roles and run with them for a period of time.  If it doesn't work, change it up at a point in time.  Even Shelton has mentioned he would like to establish roles.  It is hard to constantly try to match up, at some point you have to go with your best option for that situation.  Even if the options are average at best.  So give someone the 8th inning and someone the 9th inning and see what happens.

[ziggy]

When your best relief pitcher in 2026 was the 5th best in 2025 I'm not sure you can define roles quickly.

You can't make a silk purse out of a sows ear.

[rv78]

OR, it could just be that none of the bullpen pitchers have a defined role because none of them are good enough to fill those roles. Who is the Duran, the Jax, the Coulombe, the Varland of this bunch? No one.

[IndianaTwin]

People have clamored for the Twins to play the kids. These guys aren't necessarily young, but they are unproven, so I think it actually makes sense to move guys around to see how they respond to different situations -- empty bases, runners on, starting innings, middle of an inning, up/downs, etc. 

[karcherd]

22 minutes ago, IndianaTwin said:

People have clamored for the Twins to play the kids. These guys aren't necessarily young, but they are unproven, so I think it actually makes sense to move guys around to see how they respond to different situations -- empty bases, runners on, starting innings, middle of an inning, up/downs, etc. 

Garcia, Banda, Rogers, Sands, Topa, even Orze is in his second full season.  Where are you seeing inexperience in the bullpen, I see a bunch of washed up vets with experience.

[DataNerd]

No really going to fault Shelton for bullpen management when there really aren't any reliable options.  Bullpen is 26th in WPA, which is better than its 28th place ERA.

[tony&rodney]

Derek Shelton is working with the arms the team has in their bullpen. I can't criticize him. It's not like anyone on TD has ever has a molecule of baseball knowledge equal to Shelton. Many of us fans have experience and some knowledge but these guys are pros. 

I'm actually positive about the current coaching and manager. The player talent will improve in the next year and so will the results. Perhaps by then the Twins will have a bullpen with pitchers who have more defined roles, similar to a year ago.

[IndianaTwin]

1 hour ago, karcherd said:

Garcia, Banda, Rogers, Sands, Topa, even Orze is in his second full season.  Where are you seeing inexperience in the bullpen, I see a bunch of washed up vets with experience.

Adams (.066*), Rojas (.000), Morris (.000), Gomez (.135), Laweryson (.017), Acton (.031), Orze (.116) all entered the season with less than a year of service time. Plus SWR (1.153), Funderburk (1.122) and Sands (3.017) are still on the young side of 30. Some of those are injured or on the St. Paul Shuttle, but any who are healthy will likely pitch again for the Twins this year.

Only 65 innings have been pitched by guys over age 30. Guys like Garcia, Banda, Rogers, and Topa serve a couple of functions: A) it's a roll of the dice to see if they have something left or that can be tweaked or B) to cover innings so that the young guys aren't overexposed too early and are developed properly.

For example, in category A, Banda has been a scapegoat this year, but if you look at his game log, his entire cumulate negative WPA can be attributed to one game -- take it out and he's a net positive. I don't know whether .500 is an average "winning percentage" in WPA, but in 15 of his 22 games, he's had a positive WPA. Of his 15 earned runs, nine came in an eight-day span in early April. Of his 18 hits allowed, nine came in the same span. He's only 32 and threw 114.2 innings over the past two years with a 3.14 ERA. I've no problem with rolling the dice a little longer to see if two-plus years of quality outweighs a week of stinkage.

In category B, Garcia is 39 and is likely just here until they need a roster spot. Rogers has been pitching fine and serves as innings eater/stability factor. Topa exited today.

I'm not smart enough to pick out which ones, but several of the guys in the first paragraph will be good medium to high-leverage guys by year end or early next year. 

*Years.days of MLB service time heading into the season.   

[The Great Hambino]

If Zoll can draw relievers out of a hat, why can't Shelton?

[Bodie]

When most  of your bullpen should be riding a bus somewhere, does it really matter?

Prefer to stick it out when the 6+ERA reliever can't find the strike zone's zip code?  (Again...)  Or is it next man in before the game is utterly out of hand?  If you're trying to win the answer is obvious. 

Want more defined roles?  I see a collection that should be doing mop-up work to hopefully save thier individual careers.  At best.  Anyone see something else?

[Riverbrian]

I think the approach is perfectly sensible. 

Just wish we had batter arms to take this approach with. 

This bullpen needs some power arms... some arms that Derek can deploy when the scouting reports say that hitters 3, 4 and 5 struggle with 99 MPH heat. I like having a pre-plan for your bullpen usage. 

Our bullpen is too crafty. We are trying to locate our opponents to death. 
 

 

[Patzky]

Without hitters its moot.

[dxpavelka]

5 hours ago, karcherd said:

At some point, they need to determine roles and run with them for a period of time.  If it doesn't work, change it up at a point in time.  Even Shelton has mentioned he would like to establish roles.  It is hard to constantly try to match up, at some point you have to go with your best option for that situation.  Even if the options are average at best.  So give someone the 8th inning and someone the 9th inning and see what happens.

Sounds like that's what he is doing, just on a different level than what we might be used to.  Guess that's why they pay him a million dollars