Wednesday, March 28, 2012
Things are trending upPosted by Dan Brooks
About a year ago, sometime during John Lackey’s precipitous decline into worthlessness, I started paying attention to an odd phenomenon in which he would lose velocity during games:
Here, we can see a clear slope in fastballs thrown over the course of the game. It’s as if there’s some direct, measurable relationship between how many pitches he’s thrown and what the speed of the next pitch will be. The same is probably true of the breaking pitch, but it’s more difficult to see because the pitch mix (and speeds) are more variable.
Of course, people who look at PITCHf/x data will say that this is nothing particularly new—lots of pitchers do this—and they’d be right. But I wasn’t aware of any study describing this phenomenon in any detail. So, what I did was write a quick script that went through my database and found pitchers who had thrown at least 30 fastballs (sinkers and cutters included) in a game and had thrown 10 or more starts in either 2010 and 2011. It then fit a simple line to fastball speed by number of fastballs thrown to output a “fastball slope.”
It turns out John Lackey isn’t even the worst offender. For example, here’s a “.10” game by Jonathan Sanchez (who is the weakest link):
The “.10” here means that for every 10 fastballs, he loses a mph on his fastball over the course of the game.
“.10” isn’t as bad as it gets, though, Jonathan Sanchez is just the worst on average. Here’s a staggering “-.19” game:
Tommy Hunter starts the day blowin’ em away at nearly 96... and then barely tops 90 by the end!
Of course, the article is titled “Things are Trending Up.” How about games in which the opposite is true? For that, we need to turn to a subset of pitchers who actually gain speed over the course of their outing. Like, Justin Verlander, who does this with some sort of regularity. Here’s a “.14” game from Verlander last year:
Sure he’s got one mighty fastball in the early innings, but for the most part, he really gets cooking quite a bit later in the game, topping 100 several times.
You might ask yourself if this is a trait or just an oddity of strange games. So, I took those scores, averaged them, and then split them by year. Here’s what I found:
Each dot on this graph is a different pitcher over these two years. There’s a clear relationship between 2010 and 2011, suggesting that how a pitcher changes over the course of a game is a stable trait, likely a result of mechanics or physical attributes intrinsic to each athlete.
Dan Brooks is a Neuroscientist at Brown University. He operates BrooksBaseball.net and eats Fried Chicken during every Red Sox game, especially in September. Come follow him @brooksbaseball.