Why PITCHf/x is important

I'll admit that this is a narrow vision of why PITCHf/x is important. That's because PITCHf/x is important for a lot of different reasons. After all, there are dozens, if not hundreds, of ways that analysts have used PITCHf/x to provide deeper insights into the game that we all love to watch. The following paragraphs will outline a bunch of different ways in which PITCHf/x is important, though they're not the aspect that I'll primarily focus on with this piece. More on that later. First though, here are the non-most important reasons PITCHf/x is important.

First and foremost it helps us understand the pitches that are being thrown, where they're being thrown, and what they're doing once they leave the pitcher's hand. This is beginner stuff, but the data is very important because without it nothing more advanced could be built. Not only that, but it gives us some insight into how umpires call games, and what exactly the strikezone actually looks like. We can take that stuff even a step further as my colleague Scott Lindholm did, and look to see if umpires change how they call games as innings go by. It's not just pitchers and umpires though, we can see some pretty cool data on catchers using PITCHf/x as well. When it comes to this, nobody does it quite as well as the PITCHf/x experts Dan Brooks and Harry Pavlidis of Brooks Baseball and Baseball Prospectus. Hitters also get some action with plate discipline being a constant topic of discussion utilizing PITCHf/x data. Kevin Ruprecht and Justin Hunter give us two great examples of how PITCHf/x can be used to analyze hitters, something that gets a lot less attention than it probably should.

The most obvious and common use for PITCHf/x data is to analyse pitchers. This happens to be my specialty, as I've followed in the footsteps of former writers who have gone on to join teams and take their analyses to front offices across MLB. Mike Fast for example was the reason I got into PITCHf/x analysis, and he's gone on to greener pastures in the Houston front office. As far as analysis goes, the most basic stuff is simply having reliable velocity, movement, and location data. That allows for things like analyzing what happens when pitchers hit triple digits. That's only part of the equation - it should be combined with other statistics to get the full picture - but it's a key piece of the puzzle. PITCHf/x allows us to say that Masahiro Tanaka has the best splitter in baseball. Or maybe Hisashi Iwakuma does. We can even predict whether a pitcher will truly be an Ace, or if he's just been lucky with some deep analysis using PITCHf/x, which is really very cool if you think about it.

All of that is amazing, it truly is.

The focus of this piece though is on something that is, admittedly, a little less tangible. PITCHf/x is important because it lets us get inside the head of major league pitchers. This isn't necessarily intuitive, but it's indelibly true. You see, with each pitch thrown a pitcher makes several conscious decisions. Which pitch am I going to throw?Where am I going to try and throw it? It might help to look at a macro level example to illustrate my point. This article I wrote about Yovani Gallardo and how he's adjusted to stay relevant features some interesting facts about fastball velocity and pitch usage. It's really a psychological profile of a guy who realized that he needed to innovate or die in an industry where wisdom may come with age but performance does not. Each season Gallardo had to make conscious tweaks to his game in order to remain an effective pitcher for the Brewers. Here we're looking at things on a season-to-season basis, but the crux of my argument holds true. That is, Gallardo is making decisions that can be seen through the data that PITCHf/x makes available to us.

Obviously we can look into how much Garrett Richards' currveball drops using PITCHf/x (a lot), or how hard Aroldis Chapman really throws his fastball (hint: really really hard). But we can also raise some bigger questions like why Clayton Kershaw throws a lot fewer fastballs when he is ahead of a hitter, but Felix Hernandez throws more.

We know that Kershaw has devastating breaking balls, so his thinking might be along the lines of "I want to get ahead with my fastball so I can put them away with my breaking stuff". That's pretty standard pitching strategy, and not terribly interesting. King Felix on the other hand has the potential to be more interesting.

Felix's fastball and change-up usage increases dramatically when he gets ahead or to two strikes, so perhaps Felix likes to use the two pitches in tandem to set hitters up for a punch out. If he gets a batter 0-2, he can throw a change up low and away to try to get a called strike three (the change breaks out of the zone, but the fastball which looks just like the change stays in the lower part for the K). He could also throw the fastball just off the plate to set up the hitter, and then follow it up with a change-up that fades back into the zone for the final strike.

The other day August Fagerstrom wrote an interesting article for Fangraphs that broke down a particularly impressive at-bat from Madison Bumgarner. Now, we here at Beyond the Box Score know just how awesome Bumgarner can be at the dish. August's article focused on Bumgarner and his awesome approach at the plate, and for good reason. However he mentions something really interesting near the end:

By god, pitchers are actually adjusting to Madison Bumgarner at the plate! At the beginning of the season, Bumgarner was seeing fastballs 80% of the time. Not uncommon for a pitcher. Now, Bumgarner sees fastballs just a little more than half the time.

August Fagerstrom - Madison Bumgarner's Most Impressive At-Bat of the First Half

Right there Fagerstrom is getting inside the head of opposing pitchers! Pitchers are adjusting to Bumngarner because he's killing them at the plate. How pitchers are adjusting, and Fister's approach to that entire at-bat are just as worthy topics for discussion as Bumgarner's impressive plate discipline.

I first got into this mental aspect of pitching when I analyzed Ubaldo Jimenez a few months ago. Ubaldo has one of the most nuanced approaches in baseball, changing his repertoire and pitch mix depending on count and placement of base runners. It's obvious from studying Jimenez that he consciously changes his approach to match each situation he finds himself in. This isn't true for all though; Matt Cain doesn't change his approach no matter what the count is. Oh, there's a runner in scoring position? Matt Cain doesn't care, he maintains his approach regardless of where the base runners might be.

What is it that makes Matt Cain so different from Ubaldo Jimenez? Obviously there are differences in pitch mix, movement, velocity, etc. That's not the key difference, though. The key difference is in their approach. Approach, for the most part, is a mental exercise in which the pitcher decides how to attack a given hitter. Their approach can change, as in the case of Jimenez, depending on the situation. Then again, it may not.

I'm posing a lot of questions here without answers, but that's really the point of this article after all. This psychological aspect of the game is one of the next frontiers in analyzing baseball data. With further analysis of how pitchers approach opposing hitters, we can learn even more about the game we all love.

For the average fan, this is as simple as getting inside the head of the pros. When I was younger I always wanted to be like my favorite pros, and now we have the data to understand how they think and operate better than ever. If you're a young Oakland A's fan that loves Sean Doolittle, I can not only tell you that he throws his fastball 80% of the time, but I can show you where he throws it in the zone, and I can even show you how to grip that very pitch. For a fan it's easier than ever to get inside the mind of your favorite players, all you need is to take a gander at the PITCHf/x data we have on them.

There are other applications for this thinking as well. For example, we could start to bucket pitchers together in the hopes of identifying the mental aspects of pitching and what traits are common among pitchers. Much like how we know that Masahiro Tanaka and Hisashi Iwakuma are among a group of pitchers that rely heavily on splitters, we can group together Matt Cain and Max Scherzer as pitchers who don't change their pitch mix based on the situation.

This may seem odd or pointless--applying psychological characteristics to pitch data. In fact, it's really quite similar to something Nielsen has been doing with lifestyle data for years. Nielsen's PRIZM system assigns various psycho-graphics to groups of people based on various demographic, economic, and lifestyle data points. As an example you can go to this page and see a list of PRIZM profiles for your zip code based on a plethora of data points (some of which are included on the map and below it). Here's the number one PRIZM segment for my zip code, an example of how my demographic data is mixed with that of my neighbors to create larger groups of people.

It's not uncommon for business principles to bleed into baseball and vice versa. Jonah Keri can tell you a thing or two about how business principles helped make one of the most surprising teams in baseball history. So the idea that taking a concept like marketing segmentation and applying it to pitchers and hitters in baseball shouldn't come as shocking. Obviously, if we want to know more about Koji Uehara, it makes the most sense to look at him on a micro level using PITCHf/x. For teams, though, it might be worth grouping pitchers together because some pitchers in a group might be available via trade or free agency. In that way, identifying common traits or characteristics can be invaluable to building a team.

The first group I'd create would be the re-inventors. This group consists of guys who significantly changed the way they pitch in order to save and/or resurrect their careers. This group would theoretically include Scott Kazmir, Brandon McCarthy, R.A. Dickey, and many more. It's possible that the characteristics the Yankees saw in McCarthy that made him a trade target could be very similar to the reasons the Indians and A's took chances on Kazmir bouncing back in such a dramatic fashion.

The examples I've listed above aren't the end-all, be-all of PITCHf/x analysis. They are merely interesting parts of it that haven't been broached yet. It's easy to pull up a Brooks Baseball game log and see how hard a pitcher was throwing during a game. What takes more doing is getting inside the pitcher's head to see why exactly Daisuke threw that slider after a curveball in the first inning.

Nearly everyone got inside the head of a major league pitcher when Adam Wainwright threw a "pipe shot" to Derek Jeter in the All Star Game. PITCHf/x helped us decide if Waino really did groove a pitch or not. The grooved pitch heard 'round the world isn't the most exciting look into a pitcher's mind, after all many of thought it was coming. What it is though, is a glimpse into the more personal side of the story that PITCHf/x has been trying to tell us all along.

All statistics courtesy of Fangraphs, Brooks Baseball, and Baseball-Reference.

Jeff Long is a writer at Beyond The Box Score and Baltimore Sports and Life. You can follow him on Twitter at @BSLJeffLong.