What Went Wrong with the 2022 Twins Pitching Experiment?

[Nick Nelson]

With huge gaps to fill on a staff facing the losses of José Berríos and Kenta Maeda, Minnesota's front office took an unconventional approach to pitching this year.

Did it work? I think we can safely say: not really! But a dissection of what went wrong reveals some worthwhile nuggets to take forward.

Image courtesy of Richard Mackson-USA TODAY Sports

 

Having lost their top two starters to a deadline sell-off trade and Tommy John surgery, the Twins headed into last offseason with a barren rotation outlook. It was unclear exactly how Derek Falvey and Thad Levine would realistically be able to offset these big losses.

Yes, they had money to spend. But the next premier frontline starting pitcher to choose Minnesota in free agency would be the first. Some creativity was gonna be needed to field a contending staff, and Levine hinted as much early on.

The general manager's quotes led me to write a column around this time last year: Are the Twins About to Build a Radically Unconventional Pitching Staff?

“I think with the challenge comes opportunity,” Levine had said. “We’re going to be as creative as we can be in terms of not being necessarily hemmed into the notion of it has to look exactly the way it has always looked. We may end up looking at this from the lens of how many multi-inning guys can we add to a staff and how far does that take us?”

The Twins followed through on their foreshadowing ... to an extent. With their only stable veteran workhorses – José Berríos, Kenta Maeda, and Michael Pineda – out of the picture for 2022, the team didn't acquire proven inning-eaters to replace them. Instead, their pickups were Sonny Gray, Dylan Bundy, Chris Archer and Chris Paddack, none of whom had thrown even 140 innings the prior season.

Meanwhile, the only rotation incumbents were Joe Ryan and Bailey Ober, who had thrown a collective 120 innings in the big leagues.

The front office assembled a staff full of pretty good pitchers who were – almost uniformly – unequipped to provide any length, and so we saw Levine's vision more or less come to life: vast numbers of different arms rotating in and out to cobble together nine-inning games.

The Twins used a franchise-record 38 different pitchers. Their starters averaged 4.8 innings, second-fewest in the American League. They used six or more different pitchers in a game 31 times. 

Radically unconventional indeed. And, had this approach been successful, you wouldn't hear me complaining. But clearly it was not. The Twins ranked 19th in ERA, 19th in FIP, 20th in fWAR. Even for a club that was built around the strength of its lineup, that's not nearly good enough.

The plan, at its core, was not a terrible one: maximize the stuff of your pitchers in shorter stints, shield them from multiple trips through the order, and possibly reduce injuries from overuse. 

Alas, none of those supposed benefits came to fruition. So what went wrong, and what can we learn? Was the entire philosophy bunk, or was the execution botched? 

I would argue, probably more of the latter. There might be some merit to the concept, provided the Twins heed these lessons learned:

The starters weren't good enough, or healthy enough, even in shortened starts.

I don't dislike the idea of signing a cheap pitcher – who doesn't have the repertoire or durability to go deep – for the back of your rotation and unleashing him in highly effective 4-5 inning bursts. The problem is that this group lacked the capacity to be highly effective even with this usage.

Bundy held his own the first time through the lineup, then got mashed the second time through (.291/.327/.534), often making the third time a moot point. Archer posted an 85 ERA+ despite almost never pitching past the fourth. He placed a heavy weight on the bullpen every fifth day, and rarely left them in a good spot.

Meanwhile, the cautious management wasn't enough to prevent Ober, who only once threw even 90 pitches in a start, from being derailed by a season-ruining groin injury. It wasn't enough to prevent Gray, who grumbled about Rocco Baldelli's early hooks, from multiple significant hamstring injuries.

If the Twins want an approach like this to pay dividends, they need to find pitchers who are actually capable of excelling in shorter starts (a la Andrew Heaney) and they need to better help their players physically adapt to the altered routines.

You've got to have at least one starter who can be the workhorse.

Even with all of the above being addressed, I still think you've got to have at least one starting pitcher in your rotation who you can count on to give you some length. This strategy built around five-and-flies, piggybacking and the like becomes a lot more palatable when there is a fixture like Berríos routinely firing 6-7 innings each time through the rotation. 

That likely contributed to the decision to acquire Tyler Mahle at the deadline. He threw 180 innings in 2021 (would've led the Twins by 60), and had completed six or more frames in eight of nine starts for the Reds leading up the trade. 

Of course, Mahle proved to be the opposite of a remedy for Minnesota, and now only adds to the uncertainty of a 2023 rotation in desperate need of stable and dependable durability.

Their bullpen wasn't built adequately to handle the burden.

This is what really gets me. Levine talked about "looking at this from the lens of how many multi-inning guys can we add to a staff," and then their only bullpen addition of the offseason was Joe Smith, who could barely be counted on for one inning.

The Twins rarely carried anything resembling a long man on the staff, and would typically just march out endless one-inning relievers after short starts. This led to them frequently burning through all of their high-leverage arms on one night and burning out the back end of their bullpen for the next. 

To make a system like this work, you've got to have an array of arms capable of getting more than three outs on a regular basis. The routine of four-inning starts followed by 5-6 relievers is not a workable formula as we saw.

Losing your pitching coach mid-season doesn't help.

This one can't so much be blamed on the front office and their planning. It's difficult to anticipate such a disruptive event in the heart of your season, and Wes Johnson's abrupt departure made matters tougher as the Twins tried to hold together their experimental pitching staff through the second half.

Pete Maki undoubtedly played a significant role in architecting this year's plan, and now, as the apparent choice going forward at pitching coach, he'll be able to more directly pull the strings and execute to his preferences. 

So, to summarize...

The model of building a pitching staff with reduced emphasis on traditional 6-7 inning starters isn't bad in theory. Indeed, there's plenty of evidence that it is the inexorable direction of baseball at large. But if the Twins want to lean into this movement as they did in 2022, they need to get better at. That means:

• Filling the rotation with starters who can at least stay healthy and excel in 4-5 inning starts.
• Finding at least one workhorse type starter who can reliably give you 6+ innings each turn.
• Equipping the bullpen with enough firepower and multi-inning relievers to shoulder the load.
• Having one central mastermind oversee the operation (and if it's not working, find someone new).

As you're looking through the options available in our bullpen chapter of the Offseason Handbook, these are lessons worth keeping in mind.

View full article

[mikelink45]

You hit on it.  162 games with a minimum of 3 relief pitchers each - 486 innings of work.  No BP can handle that load.  And unfortunately that is the minimum.  After writing that I went and looked at the actual innings pitched by relievers - 635 and that is only from those on the 40 man - it does not include Winder, Duffey, and others we cut or have injured.  

Ryan had the most innings as a starter 147 - if every starter on a five man rotation matched him it would total 735 innings, but that is not the case.  Bundy was second in innings with 140, Sonny Gray had 112.  Bundy is gone, Ryan and Gray anchor the rotation - how many innings will Mahle be worth?  Can Ober hold up?  Is Maeda the pitcher we had for the very short 2020 season and can he hold up on his return? Does Paddock come back late with an innings limit?  Do we turn lose any of our rookies?

[tarheeltwinsfan]

Nick - Thank you for one of the best TD articles I have read. Great analysis. Reads like the summation of an investigation of an expert economist, hired by a company to analyze a labor problem in an industry, including recommendations for improvements. Send this to Derek Falvey and Thad Levine, for them to use as a starting point going into this off season and on into the 2023 season. 

[Game7-91]

Agree with Nick this is clearly the emerging model du jour in mlb. Makes me queasy though. On any given day, the pitching is only as good as the weakest link in the chain so to speak. I prefer length from starters, measured more by outs, less by pitch count....i have a hard time accepting 115 pitches every 5th day on average is substantially increasing risk of injury. Besides, protecting starters in this way further exposes BP to ineffectiveness and/or injury.  Regardless, this model requires starters effectively pitching at least 2 rotations through the batting order, 18 outs on average minimally, and, absolutely zero weak links in the BP. No Joe Smith's allowed. 

[Whitey333]

Excellent article.  I agree with your analysis.  Problem is this FO has done a very poor job of investing in quality pitching.  The signings and lack thereof for others really shows up.  I imagine we will see more of the same this off season.  Bargain basement pitching acquisitions usually shows up in the standings.  Hope and what ifs and potential are the words the Twins live by and sell to the fans.

[Karbo]

A couple of decent to good long guys are needed. Better coaching would help a lot also. Starters that can average at least 5 innings per start.

[Reptevia]

Like we’re finding with all aspects of this team: roster construction matters!  Way too many flyers.  How many Hyundai’s do you have to watch being towed before you buy a Toyota?

[Blyleven2011]

It's nice to have a plan implemented but if the plan is not successful you can change the plan midstream and go to plan B ... 

They did not  ...

They plan looked good at the beginning  and the twins looked like they were going to distance themselves from the rest of the division  ...

they did not  ...

the plan soured and we played below 500 the rest of the year  ...

We may of found some useful players in 2022 for 2023  ,

they did...

tap into that talent and make them better for 2023  ...

That is to be seen ...

It's my plan and I'm sticking to it  ...

[MGM4706]

Simply look at the playoffs. Very few if any of the Twins pitchers could even make the staffs of the last 4 teams standing. That's in ANY capacity!  That showed how far the Twins are from being just competitive! So if we continue down the path of getting cheap Tommy John retreads just accept the Twins as a .500 team not focused on winning. Just focused on being Profitable. That is the Pohlad mantra!

[Blyleven2011]

2 hours ago, tarheeltwinsfan said:

Nick - Thank you for one of the best TD articles I have read. Great analysis. Reads like the summation of an investigation of an expert economist, hired by a company to analyze a labor problem in an industry, including recommendations for improvements. Send this to Derek Falvey and Thad Levine, for them to use as a starting point going into this off season and on into the 2023 season. 

I agree it was a good analysis  , let's hope their plan is revised this off season and a better outcome for 2023 is on the horizon   ...

With short starts they need bullpen pitchers to throw muti innings  , plain and simple  ... 

Depth in muti inning pitchers to piggyback the  starters  ...

[Doctor Gast]

I remember very well that article you wrote last year. I also remember well the mention of some form of piggy-backing which I didn't see, which could have been a follow up article. It all made sense because of the reasons mentioned and even the pitchers the Twins were somewhat counted on never materialized (Thorpe & Dobnak as examples) not to mention Alcala. 

Piggy-backing is making use of the long relief corps. The problem wasn't that the experiment failed, the problem was the experiment (piggy-backing) was never incorporated. They continued the same old formula as if we had a very strong rotation & full supply of dominate short RPs. We had an abundant supply of profiled spot start/ long RP (hot out of ST Smeltzer, Ober, Winder, Sands and fragile arm Paddack) to begin the season.

But the fact was that we had a weak rotation & poor short relief corp. If the experiment (type of piggy-backing) was indeed incorporated and if Duffy & Pagan were not profiled as closer and their mistake were quickly reversed. That'd vastly change our outcome, but it wasn't. Injuries, depleting our pitching resourses & ineffecency were the result of over relying on a poor short relief & over stretching the rotation.

Entering '23, many of the same conditions for this hybrid pitching still exists. We don't have the money to have a stable full of work horses & an abundance of dominate RPs. You can't profile Maeda as a SP being almost 2 yrs of recouperating from TJ surgery (he was never been a work horse), hopefully Gray will be in condition after ST, Ryan will take that next step and Mahle to resemble more the SP we expected. Ober after an injured '23 shouldn't be profiled more than a spot starter/ long relief.

Although an work horse is much needed, he won't solve all our problems. We still need the long relief/ spot SP  role, especially when we have so many good candidates (Maeda, Ober, Varland, Winder, etc) most of them should be in MLB gaining experience and not wasting their time in AAA, SWR with a little bit more seasoning in AAA should be ready.

The problem with this experiment was that it wasn't incorporated, let's not make the same mistake again. 

[Otaknam]

Rocco should not be so tied to analytics. If a starter is pitching well let him continue until he is not. The Bundy and Archer retread pitcher types are the exceptions because of who they are. I’m not a fan of using a cookie cutter approach to starting pitching and ignoring intuition or gut feeling about a starter. A manager should be able to do both, like Dusty Baker. Let’s find out what some of the young starters can do. But the team has to have a very good bullpen to consistently pull the starters after five innings, a concept that Rocco seemed unaware of. 

[Jacksson]

Seems like the plan should be to have 3-4 relievers used as "openers" in every series game for 1 or 2 innings, then turn the game over to the REAL starters to chip in their 5 innings, then go back to the bullpen to close the game out.

[EGFTShaw]

Maybe I am Old School, I am Old so...

It should be a case-by-case plan, I can't help but feel you need 1 (minimum) and probably 2 pitchers who have quality starts of 6+ innings.

We also need that one pitcher who can shut down the opposition and give the team a chance to get back on the winning streak.

The last part of the season, it looked like Ryan did better when Rocco let him go more than 4.7 innings.  

Speaking of pitching coaches, how can a college team out bid a professional team for a pitching coach???

[Minny505]

I am surprised by the lack of interest, around all MLB, in not training relievers for 2 inning stints. It's well understood that every RP used in a game is raising the risk factor as all it takes is one of those guys to "not have it" that day for the perfect plan to go south. 2 inning stints reduces this risk factor significantly.

My ideal modern bullpen of 8 RP would be 4 RPs that train to pitch 2 innings per appearance and 4 RPs trained to pitch 1 inning per appearance. 

I admit that may not work in practice, but it sounds perfect to me in theory. 

[Nick Nelson]

46 minutes ago, Minny505 said:

I am surprised by the lack of interest, around all MLB, in not training relievers for 2 inning stints. It's well understood that every RP used in a game is raising the risk factor as all it takes is one of those guys to "not have it" that day for the perfect plan to go south. 2 inning stints reduces this risk factor significantly.

Right. There were so many nights where the Twins were rattling through 1-IP relievers until they got to the guy having an "off day" and things blew up. You're asking for trouble with this approach unless you have an Astros or Rays bullpen.

[Nick Nelson]

3 hours ago, mikelink45 said:

Ryan had the most innings as a starter 147 - if every starter on a five man rotation matched him it would total 735 innings, but that is not the case.  Bundy was second in innings with 140, Sonny Gray had 112. 

Right. I think the way this model is supposed to look is how Tampa Bay did it. Here are the inning totals for their starters:

• McClanahan: 166 IP
• Kluber: 164 IP
• Rasmussen: 146 IP
• Springs: 135 IP

No one was a "workhorse" and they were similarly averse to 3rd trips through the order but these guys at least lived up to the modified billing for a full starter workload.

[Doctor Gast]

50 minutes ago, Nick Nelson said:

Right. I think the way this model is supposed to look is how Tampa Bay did it. Here are the inning totals for their starters:

• McClanahan: 166 IP
• Kluber: 164 IP
• Rasmussen: 146 IP
• Springs: 135 IP

No one was a "workhorse" and they were similarly averse to 3rd trips through the order but these guys at least lived up to the modified billing for a full starter workload.

And they were relatively healthy & effective as they pitched within their profiles.

[MTV]

Let’s Look to 2023 and see who’s lined up: 
Ace: Sonny Gray, 5-6 Innings

#2: Joe Ryan, 5-6 Innings

#3: Tyler Mahle, 5-6 Innings

#4: Bailey Ober, 5 Innings

#5: Kenta Maeda, 5 Innings (bullpen maybe, returning from Tommy John)

#6: Josh Winder, likely long reliever to start the year. 3-4 Inning workloads to start, possible rotation piece further along. 
 

The Twins could get another starter and use Maeda and Winder as long relievers, or go get a long-reliever of their own, like Drew Smyly, Zach Eflin, or Matthew Boyd. Or just simply go get an ace who can eat Innings, like Shane Bieber, Carlos Rodon, Pablo Lopez, Tyler Anderson, or Chris Bassitt (all of which pitched over 175 innings in 2022)

[Original_JB]

3 hours ago, Reptevia said:

Like we’re finding with all aspects of this team: roster construction matters!  Way too many flyers.  How many Hyundai’s do you have to watch being towed before you buy a Toyota?

I don't think this is fair. You know exactly what your are getting with a Hyundai, a reliable, OK, 'everyday' car that simply starts to show it's age (in every way) in fewer years. Sonny Gray is a Hyundai. We will see, but Joe Ryan could be a Honda (I don't like the software programming of the Toyotas); Bundy and Archer are the 'salvage' titled Camaros and WRX's that have been owned by teenage drivers and look like it. Those are "Flyers".

[DocBauer]

Honestly, I'm not sure what else to add to your article, Nick. Well said sir!

I do think Bundy, by himself, as a 5th starter for a team that had to re-build almost the entire staff on the fly, wasn't a horrible choice. I mean, he kept the Twins in games and marched out there every 5th day. The problem was having Bundy AND Archer in the rotation. And the plan of Archer eventually getting stronger and stretching out just never came close to materializing. (I guess I didn't realize he was also fighting a bad hip).

I liked Gray when we traded for him, and I like him now. I didn't like ignoring the vast FA market last year and "settling" for only Bundy. HUGE mistake.

Additionally, why did it take until about August for Rocco to finally say ANYTHING about needing/wanting middle relief? Now, to be fair, Winder was doing that early until he had to slide in the rotation and then he also got hurt. So there was at least SOME attempt to set up middle relief but it blew up in a hurry! And Smith was the ONLY FA brought in along with Pagan in trade? Wow!

 

[jccracraft]

As I've said on other sites/strings. Given that the analytics say starters are pulled after going through the line-up only twice due to performance drop off, rather than fatigue, why not go back to a four man rotation? Add the extra starter types to BP with ability to go 2-3 innings?

[IndianaTwin]

1 hour ago, Original_JB said:

I don't think this is fair. You know exactly what your are getting with a Hyundai, a reliable, OK, 'everyday' car that simply starts to show it's age (in every way) in fewer years. Sonny Gray is a Hyundai. We will see, but Joe Ryan could be a Honda (I don't like the software programming of the Toyotas); Bundy and Archer are the 'salvage' titled Camaros and WRX's that have been owned by teenage drivers and look like it. Those are "Flyers".

Does that mean Ron Davis was a Yugo? 

[Original_JB]

34 minutes ago, IndianaTwin said:

Does that mean Ron Davis was a Yugo? 

I was actually going to come back and edit in a car for Pagan......

[TwinsDr2021]

3 hours ago, Doctor Gast said:

And they were relatively healthy & effective as they pitched within their profiles.

McClanahan made 28 starts which put him at 5.93 a start and averaged facing 23 batters a game (Not to far off a typical Ryan/Gray start). He didn't pitch in 2020 then went straight to the majors at age 24 from A and AA in 2019. He pitched 123.1 innings in 21. He only had what I would call a Twins start less than 6 innings with less than 3 ER's 3 times and two of them were in his last 4 starts. I would assume they were trying to keep him fresh for the playoffs. To me the big difference with him and the twins pitchers was he was WAY, WAY more efficient with his pitches. He only twice got to 100 pitches and finished at least 7 innings in 25% of his starts.

Kluber is what I believe the Twins thought all their pitchers could be, a guy that would pitch under 100 pitchers (never pitched more than 92 and rarely got to 90) and finish 6 innings when on and rarely ever pitch less than 5 (Only 5 times, 6 if you include his first start 4 2/3.)

Rasmussen was Kluber except it took him more pitches, but when his was low they seem to let him go further than Kluber.

Springs was a stud in the bullpen that was moved to a starter and given more of a leash as the season went on, never really turned into a 6 inning pitcher unless he was lights out which was about 25% of the time.

[jmlease1]

4 hours ago, Nick Nelson said:

Right. There were so many nights where the Twins were rattling through 1-IP relievers until they got to the guy having an "off day" and things blew up. You're asking for trouble with this approach unless you have an Astros or Rays bullpen.

This is what throws me as well. with an 8-man bullpen, there's no reason you can't have 2 guys designated to throw 2 innings in their relief appearances and still have enough 1 inning monsters to close out your games. the sequencing works:

• SP1 throws 5 innings
• LongRP1 throws 2 innings
• ShortRP1 throws 1 inning
• ShortRP2 throws 1 inning

Game 2:

• SP1 throws 5 innings
• ShortRP3 throws 1
• SRP4 throws 1
• SRP5 throws 1
• SRP1 comes back and throws 1 on a back to back

Game 3:

• SP3 throws 5 innings
• LongRP2 throws 2 
• SRP6 throws 1
• SRP2 throws 1

Now, for game 4 your first 2 inning guy has had 2 days off and should be ready to throw 2 innings again if needed. But you have to actually treat all your bullpen spots as being pitchers you will use, not holding 1 guy as a mop-up man or only slotting certain guys to pitch the 7th, etc.

Getting more consistent health out of the rotation is going to be hugely important; twins had 12 guys make at least 3 starts for them in 2022, but only 4 make at least 20. Bundy's best attribute last year was availability: he actually lead the team with 29 starts...but unfortunately (and somewhat predictably) they weren't above-average ones. I think they could make this work with the rotation we have (although I still favor going after a top-end guy like Rodon), but they can't be afraid to have relievers that they expect to throw 2 innings fairly consistently.

[PDX Twin]

Quote

The Twins rarely carried anything resembling a long man on the staff, and would typically just march out endless one-inning relievers after short starts.

In the good old days, there were a couple of "swing men" on the pitching staff who could do spot starts but mostly did 3-4 inning relief appearances when the starter crashed early. A 4-inning relief stint saves two 2-inning or four 1-inning appearances and keeps the bullpen fresh, especially if the game is not close.

Is this not a possible model today, where everything seems totally scripted and there is little flexibility. I thought this might be where the team was going with Winder on the roster as a 6th starter who was not actually in the rotation at the beginning of the year, but it didn't seem to work out that way.

[RpR]

They abandoned starting relievers and finishing with starters to answer the threads question.

[se7799]

If I had a dime for everytime the term "piggyback" is used on here...well I'd be a rich man.  Sounds like the "in" term to say the past four years. Though the Twins have rarely used the piggy back approach.  Oh look another dime

[Beast]

Same things that’ve gone wrong with all of their experiments.

Poor roster construction and poor management.

There’s nothing deep or philosophical, here.  There’s nothing out of their control.  The front office and management just aren’t good at their jobs relative to their peers.  It’s really that simple.