What Makes a Good Changeup?: An Investigation, Part Two

May 24, 2013

Two weeks ago, I looked at some of the factors that may impact "changeup" quality. When dealing with major-league pitchers, you are dealing with a rather select sample, so all results should be handled with care. In other words, this isn't a roapmap to pitcher development, but a single marker on the road.

Let's recap some of part one:

The faster a pitcher's fastball, the more likely he was to get whiffs with his changeup.
The difference in a pitcher's fastball and changeup velocity had a similar relationship.
Pitchers with high changeup whiff rates threw the changeup more often.
The vertical "drop" of the changeups relative to fastballs impacted the whiff rates that the pitch induced.

I thought that ground-ball rates would be less interesting than whiff rates, and left them out of part one, but the evidence suggests that I was wrong.

The mild relationship between fastball speed and changeup whiffs is retained with ground balls
There is a strong relationship between changeup velocity and ground balls—higher-velocity changeups induce more grounders.
Changeup speed doesn't impact whiff rate, so this seems like a winning move (but see the caveat in point 5).
The smaller the gap between fastball and changeup speed, the more grounders a pitcher induces on changeups.
Increasing changeup speed but not fastball speed will hurt whiff rates, a gigantic caveat to point 3.
The more sink, the better—no surprise here. Also, the more the changeup sinks relative to the fastball, the greater the ground-ball rates it induces.

Let's put those together:

Hey, velocity is good!
If you throw a hard changeup (relative to the fastball), you're trading more ground balls for fewer whiffs
Power pitchers (big fastball and a good gap to the changeup) get the best of both worlds (whiffs and grounders)
The bigger the drop, the better—sink it, and sink it more than the fastball, and you'll miss bats and get grounders.

So, we've learned that a hard sinking changeup is a good thing. Shocker. We've also learned that the amount of tail on the pitch—on its own or relative to the fastball—isn't an apparent factor. Based on this analysis so far, it would seem that being both a bat-misser and a worm-killer is a neat trick. We'd expect the pitchers with both abilities to fit the model outlined above: big fastball, good gap to the changeup in speed and drop. On the other end of the spectrum, we'd find pitchers who neither miss bats nor kill worms with their changeups. Here's where we'd expect a small pair of gaps and a general lack of velocity.

This complicated-looking chart shows ground-ball rate on the y-axis and whiff rate on the x-axis. The cross hair in the middle is the intersection of the two distributions' means. The rings are approximately one standard deviation (both rates have comparable if not nearly identical standard deviations around .088). The green line is the linear relationship between the two rates, which, as you can see, is mildly positive but not particularly powerful.

The tables below summarize the pitchers in two quadrants and include the "ring" of the chart into which they fall. The mean ground-ball rate is .505, and .295 cuts the pack in half for whiff rate. A summary of all four groups is shown before the top-right and bottom-left quadrant details. Reminder: the data covers 2011, 2012, and part of 2013.

Table Key

MPH	changeup speed
pfx	horizontal movement
pfz	vertical drop
Swing	swings per pitch
Whiff	whiffs per swing
Chase	swings per out of zone
GB	ground balls per ball in play
fMPH	fastball speed
gapMPH	fastball speed minus changeup speed
gappfz	changeup drop minus fastball drop
ring	summed z-scores of Whiff and GB rates

Averages

Group	MPH	Whiff	GB	fMPH	gapMPH	gappfz
Top-right	85.3	.38	.59	92.8	7.5	12.4
Bottom-left	83.1	.22	.44	91.2	8.1	9.6
Top-left	84.3	.24	.57	91.1	6.9	10.5
Bottom-right	82.0	.36	.44	91.4	9.4	11.7

Some extra details are included with the pitchers so you can reference the background on swing and chase rates in the first part of this series.

Top-right (the good)

First	Last	MPH	pfx	pfz	Swing	Whiff	Chase	GB	fMPH	gapMPH	gappfz	ring
Stephen	Strasburg	89	12	25	0.52	0.54	0.45	0.65	96	7.1	12.9	5
Hisashi	Iwakuma	87	9	33	0.64	0.35	0.55	0.78	91	4.1	13.2	4
Cole	Hamels	84	11	25	0.61	0.50	0.51	0.61	92	8.1	13.3	4
A.J.	Burnett	89	8	27	0.44	0.31	0.31	0.77	93	4.6	10.2	3
Jeff	Samardzija	86	7	25	0.55	0.46	0.42	0.59	96	9.5	11.4	3
Hisanori	Takahashi	81	14	29	0.58	0.43	0.52	0.62	90	8.7	13.4	3
Jonathan	Papelbon	89	11	27	0.60	0.35	0.54	0.70	95	5.7	13.4	3
J.J.	Putz	86	6	28	0.62	0.47	0.54	0.58	93	7.2	15.1	3
Ricky	Romero	85	9	34	0.50	0.41	0.39	0.62	92	7.5	14.5	3
Gio	Gonzalez	85	12	27	0.43	0.39	0.30	0.63	94	8.6	11.3	3
Hiroki	Kuroda	87	6	28	0.49	0.40	0.41	0.61	92	5.5	10.4	2
Fernando	Rodney	83	12	28	0.46	0.48	0.41	0.53	97	13.2	14.5	2
Roy	Halladay	83	9	33	0.59	0.39	0.52	0.61	92	8.5	12.6	2
Felix	Hernandez	90	7	30	0.61	0.35	0.52	0.64	93	3.6	11.0	2
Randall	Delgado	83	11	28	0.53	0.41	0.43	0.57	93	10.1	13.0	2
Kelvin	Herrera	88	13	26	0.61	0.38	0.52	0.60	99	11.6	15.3	2
Francisco	Liriano	86	14	25	0.44	0.43	0.32	0.53	93	7.5	7.8	2
Edinson	Volquez	84	9	31	0.50	0.41	0.39	0.55	94	10.1	14.9	2
Jim	Johnson	89	11	23	0.64	0.32	0.54	0.64	95	6.7	7.2	2
Tim	Lincecum	84	3	28	0.51	0.41	0.41	0.55	92	7.7	15.0	2
Zack	Greinke	86	11	27	0.47	0.32	0.37	0.64	93	7.3	13.4	2
James	Shields	85	10	26	0.58	0.37	0.48	0.58	92	7.1	12.7	2
Edward	Mujica	87	10	29	0.61	0.30	0.51	0.65	93	5.6	13.4	2
Miguel	Gonzalez	83	6	30	0.60	0.41	0.51	0.54	92	8.8	18.0	2
Felix	Doubront	86	13	29	0.50	0.39	0.39	0.55	93	7.3	14.7	2
Jon	Lester	86	14	26	0.52	0.33	0.38	0.61	93	7.4	11.0	2
Randy	Wells	82	10	28	0.50	0.40	0.40	0.53	89	6.8	9.7	1
Homer	Bailey	87	8	26	0.48	0.32	0.36	0.61	93	6.6	11.6	1
C.C.	Sabathia	87	11	22	0.56	0.34	0.40	0.59	93	6.7	7.3	1
Matt	Reynolds	83	8	31	0.61	0.37	0.47	0.55	90	6.9	15.7	1
Cristhian	Martinez	83	13	32	0.65	0.35	0.54	0.57	90	7.4	10.5	1
Johnny	Cueto	84	7	30	0.55	0.31	0.45	0.61	94	9.4	14.7	1
Joe	Blanton	84	9	24	0.56	0.36	0.45	0.54	91	7.0	10.4	1
Dillon	Gee	83	12	30	0.53	0.36	0.45	0.54	90	7.0	12.3	1
Justin	Verlander	86	12	22	0.55	0.36	0.44	0.52	95	8.8	10.9	1
Jason	Frasor	87	8	30	0.48	0.35	0.39	0.53	94	6.8	17.1	1
Josh	Beckett	88	10	25	0.49	0.32	0.36	0.56	93	4.8	9.5	1
Kyle	Kendrick	83	12	28	0.58	0.34	0.48	0.53	91	7.5	9.1	1
Jake	Peavy	84	9	29	0.51	0.30	0.39	0.57	92	7.8	13.6	1
Trevor	Cahill	81	9	35	0.51	0.32	0.40	0.54	90	8.6	10.6	1

Note: The top five of Strasburg, Iwakuma, and Hamels can be considered a group of elite off-speed pitchers. I don't think anyone would quibble with that.

Bottom-left (the bad)

First	Last	MPH	pfx	pfz	Swing	Whiff	Chase	GB	fMPH	gapMPH	gappfz	ring
Vance	Worley	84	7	22	0.42	0.10	0.26	0.35	91	6.5	5.7	4
Randy	Wolf	79	13	29	0.44	0.14	0.27	0.35	89	10.0	11.3	4
Derek	Holland	85	13	20	0.46	0.15	0.23	0.40	94	9.1	5.1	3
Guillermo	Moscoso	81	10	25	0.44	0.27	0.30	0.28	92	11.0	12.6	3
Bruce	Chen	78	11	29	0.50	0.24	0.32	0.32	87	9.0	10.5	3
Jo-Jo	Reyes	84	10	25	0.56	0.22	0.34	0.34	91	6.8	8.3	3
Josh	Tomlin	82	9	28	0.51	0.15	0.29	0.43	89	7.2	12.8	3
Jair	Jurrjens	84	9	23	0.51	0.13	0.36	0.47	90	6.0	7.5	2
Matt	Garza	86	9	20	0.46	0.22	0.30	0.38	94	8.1	7.5	2
Bronson	Arroyo	79	6	31	0.45	0.17	0.29	0.44	88	8.7	10.6	2
C.J.	Wilson	85	11	23	0.41	0.20	0.21	0.41	92	6.6	4.9	2
Aaron	Harang	83	8	24	0.44	0.14	0.27	0.48	90	7.3	9.3	2
J.A.	Happ	83	10	21	0.46	0.19	0.26	0.43	91	8.4	8.4	2
Mike	Leake	84	10	25	0.47	0.19	0.29	0.43	90	5.9	4.2	2
Nick	Blackburn	83	8	25	0.53	0.23	0.34	0.40	91	7.3	5.5	2
Jonathon	Niese	85	11	25	0.35	0.17	0.19	0.46	91	6.4	8.4	2
Bud	Norris	85	8	25	0.47	0.18	0.31	0.46	93	8.3	12.4	2
Mat	Latos	84	2	24	0.43	0.27	0.26	0.37	93	9.4	11.6	2
Chad	Billingsley	86	10	25	0.41	0.16	0.25	0.48	92	6.3	9.4	2
Colby	Lewis	83	9	23	0.46	0.25	0.33	0.39	89	6.1	9.0	2
Phil	Hughes	85	10	22	0.47	0.17	0.32	0.47	93	7.9	9.8	2
Roy	Oswalt	83	10	28	0.57	0.16	0.41	0.49	92	9.4	13.7	2
Joe	Saunders	81	13	27	0.43	0.20	0.30	0.45	90	8.5	8.5	2
Danny	Duffy	85	12	22	0.40	0.22	0.19	0.43	95	9.9	11.5	2
Luke	Hochevar	84	10	25	0.40	0.16	0.25	0.49	93	9.7	10.2	2
Barry	Zito	75	10	35	0.49	0.23	0.32	0.42	84	9.1	17.0	2
Travis	Wood	79	13	30	0.50	0.29	0.35	0.37	90	10.8	16.3	2
Brian	Matusz	83	6	19	0.49	0.22	0.27	0.44	91	7.5	7.0	2
David	Price	84	14	24	0.49	0.25	0.31	0.43	96	11.3	9.4	1
Erik	Bedard	78	10	32	0.45	0.27	0.30	0.41	91	12.5	15.7	1
Josh	Outman	81	10	28	0.44	0.28	0.23	0.40	93	12.7	14.1	1
Jordan	Lyles	82	9	27	0.49	0.28	0.31	0.40	92	9.7	10.1	1
Anthony	Bass	84	6	25	0.45	0.23	0.28	0.46	93	9.0	10.1	1
Ross	Detwiler	85	11	24	0.47	0.20	0.24	0.49	93	8.1	7.5	1
Rick	Porcello	83	11	28	0.50	0.27	0.41	0.42	92	8.9	7.9	1
Mike	Minor	85	14	23	0.56	0.28	0.38	0.41	91	6.6	8.1	1
Philip	Humber	85	9	25	0.47	0.19	0.24	0.50	91	6.3	9.3	1
Frank	Francisco	86	7	25	0.45	0.29	0.30	0.41	95	8.9	14.9	1
Jason	Marquis	81	9	30	0.44	0.21	0.32	0.50	89	8.1	7.1	1
Carlos	Villanueva	82	6	26	0.53	0.24	0.34	0.47	89	7.8	12.1	1
Jeremy	Guthrie	85	8	24	0.47	0.23	0.31	0.49	93	8.4	9.3	1
Jason	Hammel	88	10	24	0.53	0.24	0.36	0.49	94	6.2	9.0	1
Jhoulys	Chacin	85	4	25	0.47	0.23	0.30	0.50	91	6.6	6.6	1
Chris	Tillman	82	6	21	0.56	0.27	0.37	0.46	92	10.0	10.1	1
Ryan	Vogelsong	84	11	25	0.49	0.25	0.32	0.48	92	7.7	9.3	1
Scott	Diamond	84	10	21	0.48	0.23	0.30	0.50	90	5.6	5.5	1
Douglas	Fister	83	12	29	0.56	0.24	0.45	0.50	90	7.5	8.9	1
Brian	Duensing	85	14	25	0.47	0.24	0.30	0.50	92	7.5	7.9	1
Cliff	Lee	85	12	22	0.56	0.28	0.39	0.46	92	6.8	8.3	1
Tom	Gorzelanny	84	13	26	0.51	0.24	0.34	0.50	91	6.5	8.7	1
Mark	Buehrle	79	10	30	0.56	0.25	0.41	0.50	86	6.6	9.5	1
Jeff	Karstens	81	8	30	0.50	0.26	0.35	0.49	90	8.5	11.4	1
Edwin	Jackson	87	11	20	0.44	0.27	0.28	0.48	95	7.4	7.1	1
Lucas	Harrell	83	8	26	0.47	0.26	0.33	0.49	93	9.6	8.4	1
Jeff	Francis	77	8	35	0.55	0.25	0.42	0.50	86	8.3	13.2	1
Alfredo	Aceves	86	10	23	0.52	0.26	0.35	0.49	94	7.9	8.2	1
Wei-Yin	Chen	85	12	21	0.48	0.26	0.30	0.50	92	7.1	8.6	1
Kyle	Lohse	81	10	30	0.63	0.29	0.49	0.47	90	9.5	13.7	0
Jake	Arrieta	88	9	22	0.41	0.28	0.28	0.49	94	6.2	9.1	0

Note: "Bad" is a relative term, especially when referring to the guys in the lower "rings”; they're not far from average, and all of these guys are major leaguers. What's not even considered is how a changeup may set up something else (pitching backward) or whether some pitchers’ shaky changeups are even relevant.

This series was initially going to consist of three parts. To cover all the interesting details here, I might need five.

Thank you for reading

This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.

Subscribe now

Harry Pavlidis

You need to be logged in to comment. Login or Subscribe

matrueblood

5/24

Apologies if I'm missing this. Reading in https format at work, can't see the graph, might be limited in other ways. But is home-run vulnerability addressed? Intuitively, I might think firmer changeups or ones with smaller gaps would result in fewer total flies, but harder-hit flies. Maybe that's silly. Just a question. I love this piece.

Reply to matrueblood

harrypav

5/24

it's a good question, I haven't looked into that...yet.

Reply to harrypav

jroegele

5/24

Nice work Harry! I noticed that at least in spring training/opening few weeks of the season, Derek Holland's changeup had a much larger velocity differential from his fastball than in the last couple of seasons. Has he moved out of the "bad" quadrant if you look at 2013 alone? (if you care to look at it....)

Reply to jroegele

harrypav

5/24

http://brooksbaseball.net/player_cards/player_card.php?player=502706

Holland's speed gap wasn't the problem, it's the movement gap...or lack thereof. He's actually getting less of a drop gap between the hard stuff and the change in 2013, although he may have given himself a little more of a speed gap.

Reply to harrypav

rgrunder

5/24

Great article.

I know he's had a limited body of work that would show up in your data, but I was a bit surprised not to see Julio Teheran on here, good or bad. As he worked his way up, a lot of scouts thought his changeup was among the best in the minors. Now, he seems to almost never throw it. Any idea why he's not using it?

Reply to rgrunder

harrypav

5/24

I wonder if he'll work it back in after he gets more settled with his slider and sinker additions.

He didn't make the cut-off for # of pitches thrown to be included here.

Reply to harrypav

ScottBehson

5/24

Great article.

Would it be possible to look at pitchers from the recent past with hellacious change-ups- Santana, Mussina (I know this is out of your current data set, but would be a cool 6th article)

Reply to ScottBehson

harrypav

5/24

good idea!

Reply to harrypav

mnsportsguy1

5/26

This. There was something special with that Santana change, i've never seen so many awful swings off one pitcher

Reply to mnsportsguy1

mattgioia

5/24

Who's the outlier in the bottom right with the insane whiff rate and low ground ball rate?

Reply to mattgioia

harrypav

5/24

Tim Collins

Reply to harrypav

NathanAderhold

5/25

Good stuff.

I'm surprised that Jered Weaver didn't make the cutoff for # of change-ups thrown.

Reply to NathanAderhold

adenzeno

5/28

Some of those pitchers seem to have a split rather than a true change. While I know that the Splitter woarks as an off speed pitch, it does behave quite differently than a regular changeup, be it circle, palmball or whatever...Is there any accounting for that? Love this stuff. I am a HS pitching coach.

Reply to adenzeno

harrypav

5/28

a very important question. It's a layer or two deeper than the current stage, but we're headed there.
The thing with offspeed stuff is every grip is different. You can see splitters fade (Fister) or changeups tumble (Cueto). Rich Harden threw (or throws) a circle-change that acts like a slider. People think it's his splitter sometimes, too.
So my philosophy is
(a) try and identify changeup vs splitter grips as much as possible; but
(b) group them together for high level analysis

Reply to harrypav

HPJoker

8/01

So whatever happened to this?

Reply to HPJoker

harrypav

8/13

Guess what my Saber Seminar talk is about :)

Reply to harrypav

HPJoker

8/13

CAN'T WAIT

Reply to HPJoker

What Makes a Good Changeup?: An Investigation, Part Two

Thank you for reading

Latest Articles

Fantasy Four: Q&A with Chris Torres $

Five & Dive, Episode 426: Let’s Talk Turkey

The Almost-Comprehensive Fall League Prospect Rankings $

TA: The Dodgers Are Doing Five Blades $

So You’ve Decided to Trust the Robots B

Harry Pavlidis

More about:

Latest Articles

Fantasy Four: Q&A with Chris Torres $

Five & Dive, Episode 426: Let’s Talk Turkey

The Almost-Comprehensive Fall League Prospect Rankings $

Thank you for reading

Related Articles

Latest Articles

More about:

Latest Articles

Related Articles