Which Astros’ Pitchers Are Likely To Regress Up or Down from 2024 ERA results?
When we evaluate the Astros’ pitchers for the 2025 season, it’s natural to look back at their earned runs allowed per nine innings (ERA) in 2024. And, yes, that’s a starting point. But ERA, by its nature, is a “noisy” statistic. A bunch of random things can distort the ERA in a given season. Factors like the quality of defense behind the pitcher, sequencing of hits, quality of contact—and the list could go on—involve some degree of luck.
That’s why the pitchers’ 2024 ERA is a weak predictor of next season’s ERA. Sure it likely has some correlation to the next season’s ERA, but the error band around the prediction will likely be large. For instance, if the band between this year’s ERA and next year’s ERA is 1.5 runs, that’s not very useful as a predictor.
Studies have shown that some peripheral measures of ERA are better at predicting next year’s ERA than simply relying on this year’ ERA. The Statcast metric “x-ERA,” Fielding Independent Pitching (FIP), and Fangraphs’ SIERA metric are all better at predicting the next year’s ERA (as compared to relying on this year’s ERA).
Most of us realize that pitchers’ performance is subject to statistical regression, which can be either positive or negative from year to year. So, we can lump this exercise into the general category of year-to-year regression.
My objective here is to arrive at a better starting point for evaluating the Astros’ pitchers’ likely ERA in 2025 by reducing some of the noisy distortion in the actual 2024 ERA. Therefore, I adjusted the pitchers’ 2024 ERA to reflect the 2024 average of: (1) Fielding Independent Pitching; (2) x-ERA; and (3) SIERA. For most of the pitchers, the three indicators are pointing in the same direction in relation to the 2024 actual ERA.
The results for Astros’ starting pitchers is shown below. The “adjusted 2024” ERA for each pitcher is based on the average of FIP, x-ERA, and SIERA. We don’t know if the Astros will re-sign Justin Verlander, but I have included him in order to round out a 5 man rotation.
Source of data: Fangraphs Leaderboards
- Brown’s adjusted ERA is virtually the same as his actual ERA. Arrighetti and Verlander have an adjusted ERA which shows substantial improvement over the actual ERA. Valdez and Blanco show an adjusted ERA that reflects significant regression toward a higher ERA.
- After adjustment, Valdez and Brown are close to “co-Aces” at the top of the rotation. This is largely due to Valdez’s regression compared to his pitching peripherals, thereby reducing his ERA margin compared to the rest of the starting staff.
- Blanco enjoyed a surprising stellar campaign in the rotation. But some of the stellar numbers may be illusory. The difference between a 2.80 ERA and a 4.10 ERA is quite significant—-basically the difference between a very good pitcher and an average pitcher.
- As the Astros consider signing Verlander, there are a number of factors (age, for example) to evaluate. But the peripheral measures of ERA point to a major improvement over his poor end-of-season results. Maybe we won’t see great results in 2025, but the peripherals support the notion of a No. 4 or No. 5 pitcher.
- The adjusted ERA confirms our expectations of continued improvement over 2024 by Spencer Arrighetti.
I performed the same adjusted ERA comparison for Astros’ relief pitchers, as shown below.
Source data: Fangraphs
- Hader’s periperhal measures of ERA point to a significant reduction in ERA—the adjusted ERA of 2.86 is in the range of other elite closers. Pressly shows very little change from his actual ERA, which is acceptable for 7th or 8th inning work. Abreu shows an indication of future regression, but his ERA remains appropriate for 8th inning appearances. Two free agents, Ferguson and Neris, exhibit peripheral measures which point to a reduced ERA. Dubin, who likely will be the “long man” in the pen, also exhibits signs of downward regression in his ERA.
- It’s unclear whether the Astros are considering re-signing Ferguson, but if the price is right, the peripheral metrics are less than his actual ERA. Neris had a bad 2024 season with his 4.70 ERA, but the peripheral ERA measures point to probable downward movement to 3.95. Again, it’s a question of price—but one can imagine circumstances when it makes sense for the Astros to re-sign him.
- Montero’s 2024 peripherals are not encouraging. It’s unclear whether he will return this spring, but if he does, he has some proving to do.
- Scott, Ort, and King were all surprisingly good contributors to the 2024 bullpen. But this may put our expectations too high for 2025. Scott and Ort posted peripheral ERAs more than 1 run higher than their actual ERA. King’s situation is slightly better, but he also shows a significant risk of regression. I suspect that it’s unlikely all three will avoid damaging regression in 2025.
- My take? The Astros can’t count on the same run suppressing numbers from these three guys (Scott, Ort, and King). Relievers are notoriously volatile, particularly journeymen like these three pitchers. The Astros need additional relief pitchers to complete with Scott, Ort, and King this spring training. I’m sure the Astros will have some NRIs and minor league free agents to compete with them. But the outcome of this competition could turn out to be important if Scott, Ort, and King suffer negative regression.
Thoughts?