Notice: Trying to get property 'display_name' of non-object in /var/www/html/wp-content/plugins/wordpress-seo/src/generators/schema/article.php on line 52
keyboard_arrow_uptop

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:

  1. The faster a pitcher's fastball, the more likely he was to get whiffs with his changeup.
  2. The difference in a pitcher's fastball and changeup velocity had a similar relationship.
  3. Pitchers with high changeup whiff rates threw the changeup more often.
  4. 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.

  1. The mild relationship between fastball speed and changeup whiffs is retained with ground balls
  2. There is a strong relationship between changeup velocity and ground balls—higher-velocity changeups induce more grounders.
  3. Changeup speed doesn't impact whiff rate, so this seems like a winning move (but see the caveat in point 5).
  4. The smaller the gap between fastball and changeup speed, the more grounders a pitcher induces on changeups.
  5. Increasing changeup speed but not fastball speed will hurt whiff rates, a gigantic caveat to point 3.
  6. 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:

  1. Hey, velocity is good!
  2. If you throw a hard changeup (relative to the fastball), you're trading more ground balls for fewer whiffs
  3. Power pitchers (big fastball and a good gap to the changeup) get the best of both worlds (whiffs and grounders)
  4. 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
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.
harrypav
5/24
it's a good question, I haven't looked into that...yet.
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....)
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.
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?
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.
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)
harrypav
5/24
good idea!
mnsportsguy1
5/26
This. There was something special with that Santana change, i've never seen so many awful swings off one pitcher
mattgioia
5/24
Who's the outlier in the bottom right with the insane whiff rate and low ground ball rate?
harrypav
5/24
Tim Collins
NathanAderhold
5/25
Good stuff.

I'm surprised that Jered Weaver didn't make the cutoff for # of change-ups thrown.
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.
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
HPJoker
8/01
So whatever happened to this?
harrypav
8/13
Guess what my Saber Seminar talk is about :)
HPJoker
8/13
CAN'T WAIT