Oh, So NOW the Twins Want to Add to Their Bullpen

Bullpen arms are notoriously fickle things. I get the idea that the FO wants to let the season play out as they test out a few arms and see who's actually good that year, and fill in around that as the season moves on, but they do it to the extreme/poorly, in my opinion. 

They're way too slow to move on from vets that they expect to be good, but aren't performing, or who they think will magically jump another level to be a high leverage arm because their Stuff+ ratings say they will. They play the St Paul to Minneapolis option game way too frequently with young guys (while also sitting them in the pen for weeks without using them). They overemphasize years of control on the trade market leading them to give up "real" prospects for controllable assets, instead of the much easier Fulmer style trades that don't cost you anyone anywhere near the top of your system. 

I don't need them shelling out a huge contract to the Kimbrels of the world, but the Robertsons of the world should be their sweet spot. So I get the idea, and don't even disagree with not going bonkers investing in pen arms for high costs, but if it leads to a bunch of blown leads before the deadline every year, and you making what turns out to be a horrible trade for Lopez types, you need to change your strategy. Or at least get better at executing it. Their pen is almost always better in August and September under this regime, but they waste 4 months giving the Tyler Duffeys of the world every last chance while trying to get to those August and September pens. I know their whole thing is to not be hasty, but in-season adjustments need to happen before the trade deadline.

1 minute ago, USAFChief said:

Every time I read this old saw I grind my teeth.

Good relievers are NOT fickle.

Just the opposite.

Placeholders are fickle.

And filling your pen and AAA team with placeholders is what leads to problems

 

Agree to disagree. Unless by "good" you mean "elite." But even the "elite" guys are time bombs waiting to happen. 

Look at the 6 guys named in this article as offseason possible adds. Edwin Diaz is the highest paid reliever in baseball, but was absolutely brutal in his first year in NY, and just OK in his 2nd. Kenley Jansen lost his closer role in LA numerous times, and hasn't exactly been lights out this year. Craig Kimbrel was absolutely atrocious his first 2 years in Chicago, and was "eh" last year. Adam Ottavino had 1.58 and 1.45 WHIPs and 5.89 and 4.21 ERAs in his last year in NY and his year in Boston before bouncing back to a .98 WHIP and 2.06 ERA. Will Smith had a 1.4 WHIP and 3.97 ERA last year, but has a .84 WHIP and 2.84 ERA this year. David Robertson has been a model of consistency as a good, not great, reliever. So 1 out of 6. 

Banking on year to year performance from almost any reliever is a bigger bet than any other position. They're in the pen for a reason. Very few of them are just completely consistent, and those ones still tend to fall off a cliff at some point, and struggle to get back. Aroldis Chapman was the best reliever in baseball for a decade, and then he suddenly was extremely beatable for 2 years before bouncing back to some success this year. It's incredibly hard to predict year to year success for relievers. Especially as they age. And by the time they're eligible for free agency most of them are on that brink of losing a tick on the fastball, or a couple inches on the breaker, and falling apart. The elite guys get paid a ton for a reason. Trying to predict the years Ottavino types will have a 1.0 WHIP vs a 1.5 WHIP is the riskiest bet GMs make. There's a reason he only got 4 mil coming out of his NYY and BOS stints.

1 minute ago, Squirrel said:

I get that and see that. However, your chances are better you will have a stronger pen by not relying on AAA guys or AAAA reclamation projects. Yes, you do hit it right on from time to time with those types, but your chances are better if you aim higher. Imo.

I agree. It's why in my first post I mentioned the Robertson types as what should be the Twins sweet spot. Good, not great, relievers with the highest chance of being competent year to year. Snag a couple of those guys (I'd argue that's what they thought they had in Jax, Thielbar, Alcala, and Pagan for this year) to pair with your elite guys (I'd argue that's what they thought Duran and Lopez were this year) and go from there.

My problem with them is their evals of relievers. Pagan being the most obvious one. But letting Coulombe go in favor of Moran who had options left is another one. Where they take "risks" confuses me. They won't drop Gallo cuz Wallner and Larnach have options so they'd rather keep their depth, but they drop Coulombe for Moran instead of keeping both for added options. I think they thought they were coming into this year with 6 good to elite relievers, plus Moran as the 2nd lefty, and 1 long man. I think the vast majority of us saw far fewer than 6 good relievers, though.

1 hour ago, jorgenswest said:

Relievers are no more fickle than any other position. Their stat line varies more but it is random variation due to their sample size. That small sample gives the illusion that they are volatile as a group.

Teams can get in trouble trading for a reliever based on a half season sample particularly if there wasn’t success in the previous 3 or 4. Dylan Floro has ERAs that have a range that varies by almost 2 runs over the last 4 seasons. His xERA is very stable atlas’s than half a run.his xFIP is also stable with a range of 0.8. His strike out rates and walk rates are pretty stable. It is BABIP and HR rates that are not stable and that home run rate need a large sample. His HR/FB rate was unsustainably low in his good ERA season as his BABIP very high in his poor ERA season.

Find a reliever with a track record of good xFIPs, xERAs, strikeout and walk rates and you will have found a good reliever. There is the chance that the small sample remains will get bitten by a high BABIP but that isn’t in the pitcher’s control and only in the teams control as far as their willingness to put up a defense that limits hits on balls in play,

But small sample size is their job. If you can't succeed in small sample sizes you're not a good reliever. The top end of innings pitched by relievers is about 70 innings. Your job is to throw up 0s in as many of those as possible, and definitely to stay away from crooked numbers. Giving up runs in more games, or having more crooked numbers against you, year to year isn't shocking, but it's the difference between being a good reliever and a bad reliever year to year.

Yes, their sample size is absolutely what causes their stats to be volatile, but that's the job. Your bad performances often times turn wins into losses. It's a lot of pressure, and nobody is going to have an ERA of 0. But that's why it's the hardest prediction to make as a front office. Finding the guy who is on the good side of that very thin line this year vs the guy on the bad side is incredibly difficult. You can define a reliever as "good" by their expected numbers, but that doesn't make their performance good when it comes to actually winning games. Don't care if it's BABIP luck that kills you, or HR/FB rates. When you're pitching so few innings you have to be "good" almost every time. Sometimes its just not your year, and sometimes it is.

When it comes to major league baseball player predictions from year to year bullpen arms are absolutely the most volatile and fickle. And it's absolutely because of the sample size. But 5 extra "bad luck" losses this year than last year is still 5 more losses. Expected stats can give you the best educated guess you can make, but it's still a far less educated guess than any other position in baseball.

33 minutes ago, jorgenswest said:

Our understanding of sample size is very different. Pitchers don’t have control of the random variation due to sample. In a small sample pitchers can have the same skill, effectively throw the same pitches and have very different ERAs and BABIPs. They can only control how effectively they throw the ball.

I think the weight we put on "random variation" is different. I'd guess I think there's more change in how they're throwing the ball year to year than you would. And that's totally fine.

But that doesn't change that there is some pretty sizeable variation in their outcomes. And that's what a FO is trying to predict when bringing in any certain player. What will their future outcomes be? Predicting the future outcomes of relievers is significantly harder than any other position.