Did pitchers become fatigued at the end of the season?

The other day, I read a piece by Alex Speier of the Boston Globe, which explained his thought process behind how he voted in the Cy Young race. He put a lot of thought behind his vote and most of the article is more trying to figure out which stats are important and which are not. But one particular stat he noted that caught my eye was MLB wide ERA by month:

April: 3.99 (starters 4.03)

May: 4.08 (4.02)

June: 4.44 (4.52)

July: 4.39 (4.44)

August: 4.27 (4.38)

September/October: 4.42 (4.64)

You see a clear narrative here. Pitching was strong in April and May, faltered in June due to the crackdown on substances, pitchers slowly figured out how to pitch without them, but by the last month, they faltered due to fatigue. At least, that’s the obvious narrative staring us in the face. But is that what happened?

To investigate, I looked at past years. Specifically, the five years prior to 2020. For obvious reasons, I ignored 2020. In 2021, the MLB ERA average was 4.26. April and May were noticeably better, June and July were noticeably worse, and August was dead on. September/October were noticeably worse. So in theory, if there was a fatigue effect in 2021, we should not be seeing as big of a decline in September. What does 2019 say?

April: 4.35 (starters 4.33)

May: 4.46 (4.46)

June: 4.61 (4.59)

July: 4.59 (4.67)

August: 4.68 (4.66)

September/October: 4.36 (4.57)

MLB average was 4.49. New theories abound. April appears just as beneficial to pitchers as 2021. May, while technically better than the average, was barely so. Tentative explanation could be... pitchers saw a huge boost in May this year from the substances. June might just suck for pitchers. Though July being about as bad - and worse for starters - could suggest that in 2021, pitchers really did need to adjust without using substances. September supports the fatigue argument, but... August muddles things. You’re telling me there’s no fatigue effect in August in 2021 and in fact pitchers pitched considerably better, but a month later suddenly everyone is struggling? It’s possible, but I feel like this would show up in August too. That’s month five after a two month season.

Well let’s see if 2018 clears anything up:

April: 4.12 (starters 4.20)

May: 4.05 (4.15)

June: 4.05 (4.13)

July: 4.42 (4.41)

August: 4.14 (4.20)

September/October: 4.15 (4.09)

Okay now this is just feeling random. June is tied for the best month! August goes back to being a strong month. This is in general, a much more consistent year. Only July is really out of whack. Alright, I’m just going to move onto 2017 now, because I am starting to lose faith there are any conclusions to reach.

April: 4.09 (starters 4.04)

May: 4.30 (4.59)

June: 4.64 (4.76)

July: 4.36 (4.52)

August: 4.39 (4.51)

September/October: 4.36 (4.47)

MLB average is 4.45. June is back to being a garbage month for pitchers. September/October is for the third straight year a better than average month for pitchers. There might be something to the fatigue argument? April seems to clearly get a huge boost from probably the weather. I can see how May and August might see variations because maybe some years you get a hot month in May, maybe some years August isn’t as hot as normal. I can’t explain June 2018 at all though with the weather theory. Onto 2016:

April: 3.98 (starters 4.10)

May: 4.18 (4.27)

June: 4.44 (4.59)

July: 4.03 (4.26)

August: 4.33 (4.47)

September/October: 4.14 (4.30)

MLB average is 4.18. July gets to be random this time. throws clipboard 2015 what you got?

April: 3.93 (starters 4.15)

May: 3.82 (3.97)

June: 3.77 (3.98)

July: 3.85 (3.92)

August: 4.21 (4.33)

September/October: 4.14 (4.21)

......

.....

What the fuck? 2015, what are you doing? You’re breaking the few conclusions I’m able to draw here. For the first time both April and September are one of the three worst months. Is this just all random? A multitude of factors - ball change, which parks are being played at, what the weather is - all contribute to give us a truly random distribution of ERA per month?

Okay I’m going to do one final thing. I’m going to combine all five years together - and I’ll be honest, I’m just going to be averaging the five ERAs, instead of going through the process of adding up all the runs and adding up all the innings and doing it manually. The numbers are large enough that I promise you I would screw it up. Plus, the whole time-consuming aspect. But five numbers? I can do that.

April: 4.09 (starters 4.16)

May: 4.16 (4.29)

June: 4.30 (4.39)

July: 4.25 (4.36)

August: 4.35 (4.43)

September/October: 4.23 (4.33)

The average of those five years is a 4.24 ERA. I can draw one conclusion and one conclusion only from these results. I can’t even assume that the crackdown on substances affected the June numbers in 2021. We have at least one other example of a nearly identical jump in ERA in the past five full seasons of data. But I am pretty sure that pitchers in September and October, collectively, faded down the stretch. With the exception of that asshole 2015, each season either matched or bettered the average ERA of that season.

How did the Cardinals pitching fare by month?

April: 3.97 (Starters: 3.99)

May: 4.06 (3.55)

June: 4.81 (5.75)

July: 4.09 (3.44)

August: 3.30 (3.08)

September: 3.80 (4.23)

So the Cardinals as a team appear to have been pretty much unaffected. Which makes sense. This was the month with the winning streak. It does make getting established pitchers who know how to last through a long season like J.A. Happ and Jon Lester look like genius level decisions. Throw in Adam Wainwright and the Cardinals had three old pitchers who know exactly how to last through a season. I don’t know if the Cards were thinking like this or not. It’s possible they just got lucky and stumbled into this. I mean they certainly got lucky with results from Happ and Lester, neither of whom are good anymore. But something to think about.

Otherwise? I sort of feel like I presented some random, hopefully interesting stats that all amount to nothing. How the leagues fare by month is subject to too many outside factors to make any real conclusions. Plus, there have been a high amount of changes in past years. For starters, the ball has changed a couple times. September/October used to feature nearly the entire 40 man roster. The man on second rule in extra innings surely has some effect. I think the explosion of launch angle happened somewhere in this span and that managers were using starters much differently in 2015 than in 2021. So like I said, interesting, but ultimately, I don’t know what to make of these numbers.