The image above is some usage, speed and spin info on Trevor Bauer‘s 2018 season. Last season he added a slider which he threw 20% of the time to right handed hitters. This change along with some refinement of his other pitches helped him to his best season as a pitcher in MLB and had him in the running for a Cy Young award before a come-backer injured him for the final few weeks.
This off season he reported (via his excellent Twitter account) that he was working on a changeup to make himself even better. But he isn’t the only player trying to do this. If you follow baseball through spring training, which can be tough at times, you will have heard various pitchers from your team talking at about having a new pitch this off-season, or that they will be throwing some pitches more or less.
But how often do pitchers add new pitches or significantly change their pitching repertoire? We cannot assume that every pitcher is going to have a Bauer like improvement. So, what should we expect?
For this analysis I am going to be looking at pitch usage, and not at changes in velocity or spin, to determine if the pitcher is trying something new.
Thanks to PITCHf/x we have data on what type of pitches were thrown from the 2007 season onwards. Pitches are categorised into one of the following based of grip, speed and movement. Fastball (FB), Slider (SL), Cutter (CT), Curveball (CB), Changeup (CH), Split-Finger (SF) and Knuckleball (KB). Some pitches are unidentified but these currently number at less than 0.5% of all pitches.
To determine if a pitcher has added to their arsenal, firstly we need to determine how many pitches they have. To do that I have set a limit of 5% usage. If a pitcher uses a pitch less than 5% of the time it doesn’t count towards his ‘arsenal’.
To ensure a manageable sample size, and a ‘longitudinal’ element to the analysis, I limited the dataset to pitchers who had back to back seasons of at least 40IP since 2007. That gave 2,892 back to back seasons across 11 seasons, or 263 pitchers per season. I then counted the number of different pitches per season by player and compared it to the previous one and that gave the following.

Just under 75% of all pitchers don’t change their arsenal season to season. 14.3% of pitchers add a new pitch each season but 12.6% remove a pitch each season.
The fact that so many pitchers removed a pitch from their repertoire was initially a surprise until I considered all the failed starters that changed to being relievers. These pitchers wouldn’t need 5th, 4th and maybe even 3rd pitches when they pitch in relief, compared to how they pitched as a starter.
On average, per season we have seen 38 pitchers each season add a pitch. So it stands that there would be a player per team talking about adding a pitch.
But does just adding a pitch make a pitcher better?
To calculate this I compared the WAR/IP for each of the back to back season to see if there was any improvement. Armed with the difference I converted it to change in WAR for a reliever (70IP) and a starter (180IP) to see if there are any high level gains or losses from adding pitches.

On average pitchers get slightly worse each season, but pitchers who remove pitches actually get better.
Similarly, the difference between players who ‘add a pitch’ and those who don’t is so small it is insignificant. The standard deviation across 70IP is 0.77 and for 180IP it is 1.98. Even though the difference for dropping two pitches is far more significant than any of the others, this can probably be written off as small sample size ‘noise’.
As a few of you may have thought, a slight flaw in the method above is that a pitcher who went from using a pitch 4.9% of the time to using on 5.1% of the time would be credited with having a new pitch. So another method could be to count the number of pitches for each pitcher that have had a 5% increase in overall use.

This is a bit more significant. Players who don’t change their pitching percentages have a higher decrease in performance than ones who change one or more pitches. This would make some sense; you would expect that if a pitcher didn’t change their approach slightly year on year, teams and players would get used to what they throw. The difference isn’t that high so just mixing up the usage doesn’t make pitcher better automatically.
We now have had a couple weeks of games now and have had the chance to see if these new pitches actually exist and if the different approaches make any difference. These things are worth watching as the season goes on, but just don’t blindly expect more from one of your pitchers if they have a new pitch. It is more about how they use it than just having one more pitch to throw.