As Fastballs Fade, Establishing the Fastball Rides On

Last week, pitching prospect Grayson Rodriguez was asked a great question on The Baseball Barb-B-Cast. Rodgriguez is ranked third in an excellent Baltimore system, and as a player who was drafted in 2018, his tenure with the club spans both the Dan Duquette and Mike Elias eras. The question was: How has the organization changed over time?

Rodriguez started his answer with, “Everything about the organization changed but the name.” He touched on technology, pitch development, and the turnover in the coaching staff, but the part I want to focus on came right at the beginning, when he was describing the Duquette era: “Our pitching philosophy was, it was like, ‘Hey, you know, as a starter we’re going to go out in the first three innings and we’re just going to throw nothing but fastballs, and we’re going to see if that works.’ And, like, terrible. Terrible idea.”

Yup. That does indeed sound like a terrible idea. It also made me wonder whether teams are as focused on establishing the fastball as they once were. A reduction in first inning fastball rate would make sense for a couple reasons. First, fastball usage has dropped overall as teams have learned that pitchers should throw their best pitch more often, and fastballs themselves have become less effective:

Second, starting pitchers go fewer innings these days. The less likely a pitcher is to go deep into the game and face a lineup for a third time, the less sense it makes to hold something back for the late innings. Here I should admit that before I considered those reasons, establishing the fastball just struck me as the kind of received wisdom that so often gets challenged by advanced analytics. Does it work, or is that just the way it’s always been done?

To measure the potential change, I pulled data specifically for starters and compared the percentage of fastballs (specifically sinkers and four-seamers) they threw in the first inning and the fourth inning. I also included the overall percentage:

There are a few takeaways here, but the headline is that establishing the fastball is definitely alive and well. As the overall fastball rate has declined, the rate in the first inning has dropped ever so slightly faster than the rate in the fourth. Before we dive into the data, however, I’m going to be totally honest with you. At a certain point I got kind of carried away with the beauty of the graphs, so I decided to keep things going. I pulled fastball percentage for every inning just so I could watch the pretty lines curve together. After all, we’re not just FanGraphs. We’re also GraphFans:

Ok, back to the baseball. My data visualization digression turned out to be useful. It made it clear that the fourth inning is the right choice for our comparison. Fastball percentage increases in each of the first four innings, with the fifth tracking right along with the fourth. After that there’s a small gap, then the sixth through ninth track pretty closely together. Establishing the fastball is not just about the first inning, but also about the first time through the order.

It’s a bit hard to see on the charts above, but as I mentioned earlier, the gap between fastball rate in the first inning and the fourth inning is shrinking ever so slightly, especially over the last 10 years:

In 2022, the average starter threw 8% more fastballs in the first inning than they did in the fourth, but the distribution wasn’t perfectly normal. Most players were concentrated in the 4-9% range, with many more players above that range than below it. There are several players for whom the difference is so big that it should probably inform a batter’s approach against them:

There are a couple patterns on this list. First, five of the top six players throw both a sinker and a four-seamer, so it could be that they just need to get a feel for both early on. All the same, if you’re facing Aaron Nola, you want all the help you can get. It could really make a difference to know that in the first inning he’s way more likely to throw a fastball, especially a four-seamer, and that he rarely throws a changeup.

As for Triston McKenzie and Cristian Javier, you’d rarely be wrong if you sit four-seamer against them in the first inning. I have to imagine teams are aware of this, but I don’t know whether it’s actually affecting anyone’s thought process when they’re at the plate.

Cole Irvin is an interesting inclusion on the list, as he had the second-highest jump in differential from 2021 to 2022, just behind Jordan Montgomery. In 2021, Irvin threw his fastball 63% of the time in the first inning and 59% of the time in the fourth. His overall fastball rate only increased by 1%, but when he used them changed dramatically.

Irvin brings us to the second pattern, which is that a lot of these players have excellent fastballs. Kelly, Nola, McKenzie, Irvin, and Verlander all rank in the top 21 on our fastball pitch value leaderboards; Syndergaard and Javier also had excellent fastballs in 2022.

On the other hand, Morton, Archer, Wood, and Taillon all had mediocre fastballs in 2022. Did it affect their performance? The evidence certainly points that way:

Everyone but Wood was significantly worse in the first inning than they were for the rest of the game. They should strongly consider dropping their fastball usage in the first inning closer to their levels in the rest of the game.

There are also plenty of mysteries in these numbers. Over the last 15 years, pitchers throw more fastballs on the road than at home in the first inning (61.7%, compared to 61.2%), but their wOBA allowed on them is significantly lower (.353 to .378), and I can’t begin to tell you why.

Listening to Grayson Rodriguez, my first thought was that if teams started deemphasizing first-inning fastballs, some would have done so sooner than others. I pulled data for each team over the last 15 years, but I couldn’t find any such pattern. I did see a few trends: the Nationals dropping their fastball usage much later than the rest of the league, and the Phillies ramping up their first-fourth differential as they assembled a stable of multi-fastball starters like Nola, Suárez, and Zack Wheeler. But in general, inning-by-inning fastball rate is too crude a tool to let us pick up on multi-year shifts in an entire organization’s pitching philosophy. That said, there is a decent amount of year-over-year stickiness, both on the team level and the individual level, with r-values between .30 and .44, depending on what and how you’re measuring.

The last thing I did was look for a correlation between first-fourth differential and overall pitching performance. Once you adjust for league average in a given year, the correlations become so small as to be statistically insignificant, both for teams and for individual players. If establishing the fastball is hurting anyone, it’s not doing so much damage that they need to upend the way they think about pitching.