Should There Be Concern With The Recent Results Of José Berrios?

Since 2018, Jose Berrios has been one of MLB’s most reliable and consistent starting pitchers. Except for the pandemic-shortened 2020 season, he has posted 32 starts in each campaign between 2018 and 2024, and this season, Berrios is on pace for, yep, 32 starts.

In the 2018-2024 period, the fewest number of innings pitched by Berrios in a 162-game season was 2022’s 172. Concerning ERA, except for the noted 2020 campaign (56th percentile) and the awful 2022 (13th percentile), Berrios’ ERA percentile ranking has ranged between 64th and 79th, which is the ERA slot range for a number two starter. Hence, Berrios has been a consistent starter over the years. What is happening this season?

Overall, Berrios’ 2025 performance has slipped. The season began well: In his first 17 outings, he recorded a 3.26 ERA and a 1.06 HR/9 in 102 innings (six innings per start). In his nine most recent starts (from July 2 to August 17), Berrios has posted a 5.68 ERA and 5.62 FIP in 44 1/3 innings (4.92 innings per start). His 2.03 HR/9 in this period is markedly higher than the 1.24 MLB average for a starter. For the season, his 4.00 ERA ranks in the 51st percentile among starters with at least 50 innings. Mad Men’s Pete Campbell best summarized Berrios’ 2025 season to date.

Is Berrios’ 2025 campaign a tale of two parts? The first 17 starts and the next nine? Does fatigue explain much of the post-June underperformance, or are there some underlying issues that have persisted throughout the 2025 season? Let us dig into the topic.

Table 1 shows that the 2025 Berrios version is a slightly below-average starter. Except for ERA, all the 2025 numbers are worse than the MLB average for a starter. It is noteworthy that Berrios’ 2024 struggles versus right-handed hitters continue this season. His 2025 .339 wOBA is higher than both 2024’s .311 and his .297 career mark. Those elevated wOBA figures are due in part to higher home run rates compared to his pre-2024 career. For the 2016-2023 period, Berrios surrendered a home run to right-handed batters once every 35.3 at-bats. In 2024 and 2025, the comparable rates are once every 26.2 at-bats and 17.5 at-bats, respectively. The MLB average AB/HR is 29.4. Hence, based on the two most recent seasons, Berrios’ mastery over right-handed hitters is no longer.

Are there signs of in-season fatigue? If fatigue was a major contributor to Berrios’s recent underperformance, we should see a steady rise in the monthly xwOBA numbers. Why? Because fatigue could cause batters to make better and more frequent contact on Berrios’ pitches. xwOBA, more so than wOBA, can capture a change in batted ball characteristics. 

Table 2 shows that for a given pitch, the monthly xwOBA figure bounces around quite a bit. The xwOBA on his sinker was elevated in June and August, but July’s was like the first two months of the season. Berrios’ slurve’s monthly xwOBA has followed the better-than-average then worse-than-average cycle since the beginning of the 2025 campaign. The xwOBA on his four-seam fastball has exceeded the MLB average every month. The xwOBA data does not show a consistent monthly increase. Fatigue does not explain why Berrios has been below average this season, nor his poor post-June metrics. Are there other reasons that might explain Berrios’ decline? Is there a table with answers? Yes, Table 3!

Compared to 2024, as highlighted in Table 3, Berrios’ sinker, four-seam fastball, and cutter have lost velocity. His 2025 sinker has averaged 92.1 mph, 1.3 mph slower than 2024. From May onwards, the sinker’s average monthly velocity has been at least 1.1 mph slower than the 2024 offering. A 1.1 mph velocity decline from the previous season is significant. Furthermore, the average velocity of the four-seam fastball and the cutter is at least 1.2 mph slower than the 2024 comparable pitch types. That is also a considerable reduction in velocity. Notably, Berrios’ pitches have had meagre spin rate changes (less than 3%). One reason Berrios has not pitched as well as he has in the past is the sizeable decline in pitch velocity. Let us take a closer look.

FanGraphs’ website includes Pitching+, Stuff+, and Location+ data. For an explanation of these three models, please refer to this. Table 4 shows the noted Berrios data for the past five seasons. The highlights are as follows:

• Stuff+, which includes velocity and other pitch characteristics, shows the erosion of Berrios’ stuff since 2021. In 2021, his sinker’s Stuff+ was 110. For context, Jacob deGrom’s 2025 sinker has a 110 Stuff+ score. That is a good sinker.
• However, the Stuff+ score for Berrios’s sinker has steadily declined, and his 2025 Stuff+ grade is 89. The Stuff+ scores on all his pitches have declined since the 2021 campaign. Overall, his 2025 Stuff+ score is 89, which is below average.
• One consistent Berrios attribute is that he locates his pitches well. His 102 Location+ score is better than average. All his 2025 pitches are either a tick below average (99) or are better than average. Notably, the Berrios changeup’s 112 Location+ score surpasses the 111 posted this season by some fella named Paul Skenes.
• Back to the negative. Pitching+, which FanGraphs’ Owen McGrattan described as an assessment of “the overall quality of the pitcher’s process,” does not show Berrios in a good light.
• Each of Berrios’ pitches is below average according to their respective Pitching+ scores. Overall, Berrios’ 93 grade is sub-par. Furthermore, since 2021, the collective quality of Berrios’ offerings has deteriorated.

Okay, the quality of Berrios' 2025 pitches is below average. Yet, how does one explain the first 17 starts that generated a 3.26 ERA, and the next nine outings yielded a 5.68 ERA? There are two answers: defence and regression. 

High-quality defence should suppress a pitcher’s ERA. Before July 1, Toronto provided excellent defence behind Berrios, as evidenced by the nine Fielding Run Value (“FRV”) score. However, after June 30, Toronto’s defence was below average when Berrios was on the mound (-3 FRV), which would put upward pressure on a pitcher’s ERA. Please note that Baseball Savant does not provide FRV pitcher-specific data by batter handedness.

The second explanation of Berrios’ poor box score results after June is regression. Put another way, Lady Luck looked favourably upon Berrios before July but has since cast an unfavourable eye upon him. Berrios’ pre-July xwOBA numbers are remarkably like the post-June scores. However, the xwOBA-wOBA deltas are pretty different. For example, before July, Berrios’ overall .343 xwOBA exceeded his .303 wOBA by 40 points. However, after June, his .344 xwOBA was surpassed by the .373 wOBA. That is a 69-point swing, which is significant. A similar pattern exists concerning the pre-July and post-June splits by batter handedness. Therefore, because the pre-July and post-June expected results are similar, Berrios is pitching the same throughout 2025. What changed after June are the box-score results (ERA, wOBA, etc.), partly due to regression/luck.

Indeed, other factors may have contributed to Berrios’ post-June underperformance. However, the deterioration of defensive support and regression are among the factors.

What does all of this mean for the balance of the 2025 regular season and postseason? Berrios’ underlying metrics (xwOBA, Pitching+, and others) are not easily improved in-season. Unless Lady Luck switches course or Berrios takes pitch sequencing to another level, Berrios is likely to perform at a below-average level for the rest of 2025. If one defines a number three starter as average, then Berrios projects to be a number three or four starter for the balance of this season. Furthermore, based on 2025 performance, Berrios is a probable candidate to be removed by John Schneider from the postseason starting rotation. 

The Last Word
Concerning the number of starts, innings, and ERA-percentile rankings, Berrios has been a remarkably consistent pitcher during his career. However, Berrios has struggled this season. His stuff has deteriorated from previous levels, and his underlying metrics (xERA, FIP, and others) are below-average. Furthermore, the underlying metrics were below average before his recent (post-June) box-score struggles. Berrios’ pre-July 3.26 ERA has risen to 4.00 (as of August 17), not because of fatigue but due to below-average defensive support and regression/reversal of good luck. In summary, if one expects second starter performance from Berrios during the balance of 2025, there are reasons to be concerned.