Decisions, Decisions: The Bullpen
Where penny pinching has real consequences
Bear with me as I take you through the bullpen from a different angle, with different numbers, some you may have never heard of and some you certainly don’t see every day.
I’m concerned about the bullpen because it wasn’t what it once was and appears to be an area where Dana Brown is willing to save a few bucks, which likely means more one-appearance stints and constant traffic back and forth from Sugar Land and/or the waiver wire.
When it happened, I mentioned that losing Hector Neris was a bigger deal than some realized, but then came Phil Maton and Ryne Stanek.
None of them were perfect, but somehow, together, they made it work, though my blood pressure was worse for the wear at times.
Then last offseason, the brass decided to fix a problem that didn’t exist - closer - and, as we’ll see shortly, created unintended consequences.
That fix - and Josh Hader’s resulting salary - is probably why Brown has hinted at the bullpen being the place to squeeze a buck for the good of the bottom line.
The Numbers
This is where the different numbers come in. You won’t find ERA or xBA or HardHit% here.
A quick explainer:
The 2nd, 3rd and 4th columns from the left tell you how many games the reliever added positive or negative Win Probability when he pitched and then the percentage of games in which his NET WPA was positive.
Next is the reliever’s Net WPA for the season.
Shutdowns and Meltdowns are one of my favorite off-the-radar stats, because these are basically SUPER WPA added or subtracted in an appearance (6% WPA added or subtracted).
Finally, inherited runners and inherited runners that scored.
These numbers do not include starts (Dubin and Scott) or games where the WPA was neutral (0.00) for an appearance (there were several).
For context, the MLB average IRS was 32.9%, making the average strand rate 67.1%.
Abreu and Hader were used too often, as was Scott before he was injured and Pressly likely would have been if he didn’t spend time on the IL. The season necessitated that the way it played out, but there simply have to be more options, even if just occasionally, in higher leverage situations. This is an area I believe Neris could have helped if he had been available all season.
Bryan Abreu is good, if not always smooth or efficient. 12 meltdowns seem like a lot, but Abreu added win probability almost 80% of the time he pitched, which outside of Hader was the best on the team. It’s sometimes terrifying and angina-inducing, but he gets it done (mostly).
Those unintended consequences mentioned above? Ryan Pressly had a bad season after being relegated to the set-up role. Can he adjust in 2025 or have the Astros ruined one of the best postseason closers of all time?
Hader mostly overcame his shaky start, but we’ll see if he can fix his home run problem, which was three times his 2023 HR/9, but below his 2019 number. Just one of those years or something to be concerned about for the next four seasons?
I like King and Ort, but they were used almost exclusively in low-leverage situations, so it’s difficult to come to any grand conclusions.
I’m not a Caleb Ferguson fan and these numbers help make my case (he is also a free agent).
Seth Martinez was up and down, but other than a fantastic 85% strand rate, his numbers were mediocre at best.
By the counting stats and traditional numbers Tayler Scott had a fantastic season, but these numbers and the chart I’m about to share hint that maybe we need to slow the roll a bit.
Shawn Dubin pitched well for the role he was assigned. A long reliever with a plus WPA and who stops the bleeding by stranding 78% of runners is a good thing.
The Chart
Hader, Pressly and Abreu are a given at this point, so I didn’t include them in this part of the exercise.
Admittedly, the benchmarks I use below are arbitrary and of my choosing.
That said, it’s meant to help clarify all of the numbers above and you’re free to choose other criteria if you don’t like mine.
My takeaways:
Almost everyone was good at adding WPA on a per-appearance basis.
Most were good at adding WPA for the season, but this is where Scott fails his first test.
It felt like too many meltdowns as a group as the season happened, but the bullpen as a whole had very similar shutdown/meltdown numbers compared to 2023 (also failure 2 for Scott).
Everyone but Scott allowed less than 33% of inherited runners to score (failure 3 for Scott).
To be clear, I’m not saying Scott was “bad”, rather he had really good “traditional” numbers (such as ERA), but little effect on the team winning as a whole.
Conclusion: Talk to Me in March
A lot will change between now and the spring, but my personal preference would be: Neris, King, Scott, Dubin, and Martinez, though that would leave King as the only lefty.
Brown’s comments make me think that Neris won’t return, despite hitting every benchmark above.
Neris signed a $9,000,000 contract with the Cubs last offseason, but was miscast as a closer, which made him available down the stretch for the Astros at the minimum.
I doubt another team will make that mistake, but I also think Neris will find more than the Astros are willing to pay.
That would open up a spot, but Ferguson is a free agent who is also likely to find a better deal elsewhere.
The goal here was not to attempt to set the bullpen for next season, that’s almost impossible with what we know at this point.
The goal was to dig past the traditional numbers and add perspective on the performances from last season using alternate metrics and provide food for thought/discussion.
Thank you for reading!
