Are there any patterns in the profiles or traits of major league pitchers that repeat performance with this KPI from one year to the next, that can help us better understand when a pitcher is more likely to repeat performance?

I’m particularly interested in Pitcher Hard-Hit rate, a metric that is not stable year-over-year. 

Just because a pitcher gives up a high volume of hard-hit balls in year 1, does not necessarily mean they will give up a similar rate in year 2. There is a good chance we see that pitcher regress to the mean. But what about the pitchers that do give up a similar Hard-Hit% in year 2 as they did in year 1? 

I decided to look into pitchers who perform poorly when it comes to hard-hit balls allowed in year 1, and what separates the pitchers who then go on to bounce back and significantly improve performance the following year, versus those who repeat their struggles.

Starting by comparing exactly that, pitchers from 2021-2025 who had a 41.5% or higher Hard-Hit% rate (roughly 30th percentile in any given year) in year 1, where one group is at that level or worse in year 2, and the other group improves by a substantial amount (surpassing 50th percentile) and comparing means of a handful of KPIs between the two groups.

The KPI presenting the most interesting difference is the In-Zone Whiff% advantage group 2 has over group 1. In-Zone Whiff% is typically a good indicator of good arsenal and individual pitch stuff numbers, and good stuff has a strong relationship with the ability to induce weak contact as well as overall performance. So it does not come as a huge surprise to me that we see pitchers with improved performance between years performing better, on average, in this area.

And what about the other pitchers who don’t land in that bottom 30th percentile? What might make a pitcher consistent from one year to the next, regardless of where they ranked, versus the guys that not only go from bad to not bad, but also those that go from mediocre to good, or good to great?

I then split pitchers into 2 groups with the following criteria:

• Hard-Hit% persistence - year 2 within 10 percentile (higher or lower) of year 1
• Meaningful improvement - year 2 greater than 20 percentile higher than year1

Only Zone% and Chase% returned non-zero effects in predicting meaningful improvement when fitting a regularized logistic regression. Chase%, interestingly, returned a negative coefficient signaling better chase rates made a pitcher less likely to make meaningful improvement year-over-year. However, the magnitude of the effects for both Zone% and Chase% were very weak, and showed low predictability in determining year-over-year improvement. And while I wouldn’t expect the ability to directly predict Hard-Hit% improvement via this set of KPIs, or any KPIs for that matter, we would be more interested in at least a bit of trend or pattern, which we are not getting here.