To celebrate this back-to-school time of year, I’d like to conduct
a tutorial of sorts. What I want to cover is how to identify the
pitches of any pitcher’s repertoire, given the fabulous pitch-f/x data
target="new">several others and I have been working with this
season. What does he throw?
How hard does he throw each kind of pitch? What sort of movement is he
getting? These questions can be answered for any pitcher, if we’ve
collected data for a few hundred of his pitches. Much of what I’m going to say today has been said elsewhere, but
I think it’s useful to gather the basics in one place.
The pitch-f/x data stream has many variables, but I believe three are
sufficient for determining pitch types: pitch movement (horizontal and
vertical) and pitch speed. I’ve discussed these quantities in two
target="new">previous target="new">pieces—refer to those articles for more details. I have found the best way to visualize
different pitch types is to plot pitches on a two-dimensional scatterplot, with the vertical
and horizontal axes representing the pitch movement. The third
variable, pitch speed, is indicated by the color of each point. As
always, the view is that of the catcher, so a pitch breaking away from
a right-handed batter will have a positive horizontal movement.
The graphic below shows such a scatterplot for about 1,000
pitches thrown by a major league pitcher. The legend on the lower left shows how
the color coding is related to pitch speed.
easy to see four distinct pitch types, so let’s take this one pitch type at a time.
Fastball—This pitcher generally throws his fastball in the
low-90s, but he occasionally humps it up over 95 mph, as you can
see by the few purple pitches in the area labeled “Fastballs.” As for
movement, we see these fastballs have vertical movement
of 10-15 inches. This does not mean that the pitch rises in an
absolute sense, but rather arrives at the plate around 10-15 inches
higher than a theoretical pitch without spin. In reality, the force of
gravity would cause a hypothetical spinless
fastball to fall about 30 inches as it travels to home
plate: an actual fastball falls less than that because of the backspin
imparted to it by the pitcher.
Another thing to understand is the horizontal movement: here we see
these fastballs are moving to the left (i.e. towards a right-handed
batter) on average about four inches. That’s not a small amount; it’s
only a little less than the horizontal break on the curveball.
This fastball lateral movement is almost
universal: essentially every fastball is a “tailing” fastball. Some
move more and some move less, but very, very few pitchers throw a
fastball that doesn’t tail, at least to some degree. This is something
I had no idea of when I started looking at the pitch-f/x data.
Change-up — The change-up is supposed to look like a
fastball coming in, but of course it’s thrown with less speed, so the
batter has trouble timing it. The cluster of points labeled
“change-ups” in the above graphic fits this description. The speed is in the
low 80s, from 10 to 15 mph slower than the
fastball. Furthermore, this change-up moves rather like the fastball:
it tails in a bit more and it “hops” a bit less, but it’s certainly a
close cousin of the fastball, as far as movement goes.
The observant fan of pitching will now be thinking: “Wait, can this be
right? The good change-ups that I see tend to dive near the
plate, they ‘fall off the table’, as they say. I don’t think they
have the vertical movement of a fastball.” Well, you’re right. And me,
too—we’re both right. The change does “fall off the table,” but
that is thanks to gravity and not to any particular trickery applied
by the hurler—the change simply takes longer to get to the
plate, allowing gravity to do its handiwork for a bit longer.
By the way, the movement on a change-up will vary a bit with respect
to a pitcher’s fastball. Sometimes they have almost exactly the same
vertical and horizontal movement as the fastball, but often they will
have slightly less “hop,” as we shall see below.
Curveball—The curveball is the easiest pitch to pick
out, both when watching on TV or when looking at pitch-f/x data. Heck,
you can even ID a curveball when sitting in the stands of the
ballpark, even without checking the radar gun. Generally, the curveball will be a
pitchers slowest pitch: this one is thrown in the seventies, quite a
bit slower than the other pitches. It is also the only pitch, usually,
that has negative vertical movement. Although the size of the
vertical break may be modest — 4-6 inches in this case—the downward break coupled with the pull of gravity on this slow
pitch, gives the impression of a ball rolling off a table.
The horizontal movement can vary quite a bit from pitcher to pitcher
(as can the vertical break for that matter). But, generally, we see
curveballs breaking down and to the right for a right-handed pitcher.
If you’re curious about what a lefty’s curve looks like (or any
lefty’s pitch, for that matter), simply look at the above plot in a
Slider—The slider falls somewhere between the fastball
and the curve, both in terms of speed and movement. Indeed, our
pitcher’s slider is usually thrown in the high 80s—slower than
the fastball, but zippier than the curve (and the change). Again,
there is a whole spectrum of sliders out there: some are closer to the
fastball, some morph gradually into curveball territory.
Some sliders, as we’ll see, show no horizontal or vertical break
whatsoever. Since it’s the ball’s spin which gives it break and since
it’s hard to imagine throwing a slider without spin, this seems like a
contradiction. However, if the axis of spin is aligned with the
direction of the pitch (like a football toss, or a … gyroball),
there will be no break. I believe this is what we are seeing when we
see a pitch with no movement. I wouldn’t call these gyroballs, by the way. Or maybe I should—in any case, a
number of different pitchers throw sliders that fit the description, and I don’t believe it’s anything new.
So, those are the four most basic pitches, but of course there are
many, many variations: splitters, cutters, 12-6 curves vs. sweeping
curveballs, underhand deliveries, knuckleballs and lots more. Instead
of taking it one pitch at a time, I’m just going to show some plots
for selected pitchers. I’ll try to identify these variants as we go. If you refer back to the above plot (which belongs to
Kyle Davies, by the way), I’m sure you’ll be able to identify the most of the pitches we’ll be seeing.
Here are the movement/speed plots for three pretty good major league
pitchers. Look how similar their repertoires are:
Each of these guys relies on a good fastball, a change-up and nasty
curveball. None of them seem to be throwing any sliders.
Bedard is a lefty, of course, which is why his scatterplot
looks flipped around. Actually, based on just these plots, Burnett
seems the most formidable pitcher: more of his pitches top 95 mph, his
fastball and change-up (which I believe is actually a splitter) seem
to have more sink and his curveball has a bit more break and it’s
thrown harder (mostly above 80 mph).
This might be a good time to point out that there is more to being an
effective pitcher than what is shown in these plots. Much more. What
is left out, of course, is command. You can’t tell, from these plots,
if these pitches were strikes or balls, if batters swung and missed or
if they were all wild pitches. These plots show us what kind of
pitches each pitcher throws, but not how well he throws them.
Since I mentioned Burnett’s splitter, let’s have a look at some other
guys who are known to throw this pitch:
Accardo’s splitter, the red cluster hanging off of the fastball blob,
really seems to dive quite a bit, which is of course, the primary
characteristic of this pitch. Note that its vertical movement sits
around zero (or even below), which is about 3-5 inches more drop than
Beckett or Bedard achieve with their change-ups.
Actually, despite the original name, split-fingered fastball, the
splitter really works like a change-up—it’s thrown at
change-up speed (typically in the low 80s) and is meant to look like a
normal fastball coming in. The splitter can be more effective than the
straight change, due to its greater sink, but I believe it’s more
difficult to control. Schilling’s splitter, as you can see in the
plot, is not very well defined, and in fact, overlaps quite a bit with
his sliders (see the huge red blob in the middle of the plot). Schilling himself href="http://38pitches.com/2007/04/09/3-3-and-headed-home-4907-vs-texas/"
target="new">noted the difficulty of controlling the splitter:
Sammy [Sosa] goes to 1-2 and I throw a ‘splider’ (which is what I call my
split when it doesn’t drop, but instead cuts hard from right to
left. I also have a ‘splinker’, which is the reverse of the
splider. If I could master either of them I’d throw them 90 times a
game), he tips the ball and Tek hangs on for strike three.
Frasor’s splitter really dives a lot compared to his fastball, which would make you think it’s a very tough pitch to hit, but he
isn’t throwing it often. Perhaps he’s having trouble controlling it.
Some pitchers’ movement plots just look weird—until you realize that
these guys throw underhand or sidearm. Pitches thrown by these
eccentrics have a different spin axis, so the movement is very different.
Throwing sidearm definitely limits the variety of pitches you can
throw. Basically, most sidearmers throw only a fastball and a
slider — the whimsically named frisbee slider.
Bradford drops down very low, which gives his fastball topspin,
instead of the usual backspin imparted by the overhand motion. The
red/orange points in the plot are Bradford’s “fastball” (thrown around
80 mph) and they move quite a bit like left-hander’s curveball, don’t
they? (See Bedard’s plot, above.) It’s curious that sidearmers like
Bradford generally have horrific platoon splits (Bradford has a
very hard time with left-handed batters), even though their fastball
looks a like like a good left-handed pitcher’s curveball.
Meredith is an exception to sidewinders throwing only two pitches
— you can see his change-up thrown with the same movement has
his fastball, but 10 mph slower.
The next plots show the repertoires of three closers. Papelbon and
Putz have rather typical closer profiles: they throw hard (many blue
points, meaning over 95 mph), mostly fastballs, with a splitter as the
primary off-speed pitch. Putz seems to throw a few sliders, as well.
In contrast, Mariano Rivera is a completely different animal. Rivera’s
celebrated cut fastball makes its appearance in this plot. In
fact, almost every pitch in this plot is a cutter thrown between 90
and 95 mph. As you can see, this variant of the fastball moves to the right a bit, but
generally maintains its vertical movement. I think the difference
between a slider and a cutter is really one of degree—if you
gradually increase the “cut” on a cutter, you will end up throwing a
The straight change-up is a pitch you gotta love. Announcers
often hype up the “power vs. power” matchup of the fastball pitcher
against the fearsome slugger and seeing a mighty swing-and-miss on a
high heater is an exciting moment, it can’t be denied.
But I prefer
to see the heavily-muscled behemoth fooled by a slow ball, watch him as his bat whips
through the strike zone when the pitch is still a couple feet from the
Here are the plots for four of the best change-up artists out
there. Gagne’s change (the large red cluster of pitches) has quite a
bit of downward movement, relative to his fastball, while Hamels‘ and
Santana’s slow ones (large red clusters, again) are only slighly lower
than their fastballs.
Hoffman’s celebrated change-up is the slowest
(under 80 mph), although his fastball is also slower than the
others. What really counts is the difference in speed between the
fasty and the change. It would be interesting to go through the data
and see who has the largest difference.
Well, by now you’re probably starting to feel like you’re in analysis and
your shrink has asked you to look at too many ink-blot diagrams. I
certainly do, and although I’m tempted to sneak another graphic or two in
here, I’m going to resist. However, if you’re curious about other
pitchers, see the Resources section below, where you’ll find pitch
charts like these for a whole bunch of guys.
References & Resources
I have created these plots for about 150 pitchers. The following pdf files are approximately 2 Mbytes each: