keyboard_arrow_uptop

What's an average fastball speed these days? Which pitch types get the most groundballs? It's been several years since Harry Pavlidis last looked at the benchmarks of each pitch type using PITCHf/x. Let's revive the tradition.

Since 2011, the Pitch Info data set has improved in several ways. Harry, Dan Brooks, Lucas Apostoleris, and I, along with readers like you, have vetted our classifications many times over. The data is now corrected for system calibration errors. We've been able to spot more pitchers' split-finger fastballs, moving them out of the changeup group and correctly into the much smaller splitter category. We've also been able to identify pitch sub-types that add to the depth of research we can do, such as by creating a separate “slow curveball" tag for pitchers who throw eephus-like curves in addition to a normal one.

I've broken this analysis up to check on a few aspects of pitch attributes. The data below represents all pitches thrown in baseball in the 2014 season. The horizontal movement of left-handed pitchers is mapped across the Y-axis, so the charts below represent all pitches, but are presented as though coming only from right-handed pitchers. I've also excluded screwballs, because in 2014 almost all the pitches we labeled as screwballs were Trevor Bauer's “reverse slider." If you're really curious what that pitch is like, I invite you to visit his Brooks Baseball player card.

Physical characteristics

Pitch type

Speed

H-mov

V-mov

Dragless H-mov

Dragless V-mov + gravity

Four-seam

93.2

-4.8

9.3

-6.6

-14.5

Sinker

92.0

-8.5

5.8

-12.4

-20.7

Cutter

88.9

0.6

5.6

2.0

-22.8

Splitter

84.8

-6.1

2.6

-8.8

-30.8

Slider

84.5

2.6

1.0

-5.1

-33.3

Changeup

84.1

-7.8

4.7

-11.4

-28.2

Curveball

78.4

5.1

-5.7

8.9

-49.6

Knuckleball

76.3

-0.2

1.5

0.0

-40.6

Slow Curveball

72.2

7.2

-8.5

11.5

-62.0


A few thoughts:

  1. Unsurprisingly, four-seam fastballs are the fastest pitch type, coming in at an average of 93.2 mph. In 2008 it was 91.8 mph. We've gained a mile and a half in just six years.

  2. Sinkers and cutters—especially hard cutters—are essentially mirror images of each other to a hitter, as they are often designed to be. They get similar amounts of drop at a similar speed, but break in different directions.

  3. Knuckleballs have a slight bias toward backspin, but in terms of horizontal movement there is no pattern whatsoever—as the pitch is designed.

  4. In general, changeups tail more and drop less than splitters. Changeups' spin pattern and path to the plate imitate a slow sinking fastball, while splitters are more like four-seamers with the bottom dropping out.

  5. Slow curveballs have more spin movement than regular curveballs, which means one (or both) of the following things: (a) there's a selection bias of slow curveball pitchers toward those who get lots of RPM on their curve normally, and/or (b) the throwing motion of the slow curve lets pitchers add spin to their normal curve.

Pitch outcomes

Pitch type

Swing %

GB %

LD %

FB %

PU %

AVG

SLG

ISO

Whiff/Swing %

Four-seam

45.0

36.5

26.4

28.0

9.1

.264

.424

.161

18.0

Sinker

43.9

53.8

24.2

17.8

4.2

.293

.424

.132

13.2

Cutter

48.5

45.8

24.4

22.6

7.3

.261

.395

.134

21.2

Splitter

53.8

55.1

22.4

17.5

5.0

.208

.321

.113

34.3

Slider

48.3

45.6

24.1

22.4

7.9

.212

.327

.115

34.8

Changeup

51.6

50.6

22.9

20.3

6.2

.239

.373

.134

31.5

Curveball

39.7

50.4

24.0

20.4

5.3

.219

.328

.109

31.1

Knuckleball

47.0

46.0

22.2

22.7

9.2

.228

.375

.146

25.1

Slow Curveball

40.4

51.6

21.7

22.4

4.4

.197

.279

.082

23.7

Some more thoughts:

  1. There's a huge difference in the batted-ball types between four-seam and two-seam fastballs. Part of that surely is related to the locations those pitches are thrown to, but it's a big gap.

  2. Fastballs, both two-seam and four-seam, get hit fairly hard and aren't whiffed on very often. Part of that is probably because of their relatively frequent usage, and also because of what counts they tend to be thrown in.

  3. Splitters are very good at getting swings, and very good at getting whiffs on those swings. They're also the best at inducing grounders and have the second-lowest slugging percentage allowed. A well-executed splitter is a weapon unlike any other.

  4. Slow curves are unique in that they don't get many whiffs but they do appear to get very poor contact.

Fastball speed benchmarks by role

Overall

RHP

LHP

Starter

92.7

93.0

91.7

Reliever

94.0

94.2

93.4

Presenting the precise benchmarks for each pitch attribute for right-handed and left-handed pitchers and starters and relievers might be data overload. To give you a picture of one instance where there's a very visible difference, fastball speed is 2.5 mph higher for a right-handed reliever than for a left-handed starter. Adjust your expectations by pitcher accordingly.

The fastball-slider continuum

Pitch type

Speed

Whiff/swing %

GB %

Four-seam

93.2

18.0

36.5

Hard cutter

90.1

19.8

45.8

Soft cutter

87.5

22.5

46.1

Slider

84.4

34.5

45.6

Based on the research I did a few weeks ago about the distinct attributes of fastball-like and slider-like cutters, here's a look at a few traits of the pitches on the fastball-to-slider spectrum. What's interesting is that there's not a consistent change from pitch to pitch. The drop-off in speed is linear, sliders get way more whiffs than even the soft cutters, and only four-seamers get hit in the air much more than average.

A look back at knuckle curveballs
A little over a year ago, I found that knuckle and spike curveballs are thrown harder and generate different batted-ball tendencies. Let's see if it held true in 2014.

Pitch type

Speed

Whiff/swing %

GB %

Knuckle curve

80.0

32.7

55.4

Regular curve

77.8

30.5

48.5

Sure enough, the trends are the same. I should add that at least 89 pitchers threw a knuckle curve in 2014.

I encourage you to use the tables above as reference when you want to know the physical and statistical attributes of each pitch type.

You need to be logged in to comment. Login or Subscribe
Tarakas
6/19
Thanks for all of this work. Very interesting!
padresprof
6/19
Thank you for properly labelling pitch speed as speed and not velocity.
skuddler
6/20
Man. Splitters seem to produce amazing results. The get the most swings out of hitters and hitters miss them the most often. They also generate the highest ground ball rate and the second lowest line drive rate.
wjmcknight37
6/20
Great stuff. I was recently thinking I wanted to track down this information and here it is! Thanks.
ravenight
6/22
Great stuff! One question: what's the typical release distance for pitches? Is it the same for all pitches or does it vary? I assume that it varies per pitcher, but what's the range of that variance? I ask because the effect of the dragless gravity adjustment seems to be pretty different for different pitch types - slow curves have [V_Mov] - [Dragless V_Mov + Gravity] = 53.5, whereas for 4-seamers it's 23.8. Some back-of-the-envelope calculations imply a release distance of ~51 feet for the slow curve vs. ~38 feet for the 4-seamer, so clearly I'm leaving something out (probably an initial downward velocity produced by the arm motion, release snap, and plane of the pitch) so I'm wondering how the adjustment is calculated (does Pitch/fx know the release distance of each pitch?).
ravenight
6/22
Ah nevermind, I see that Brooks Baseball uses 55' as the release distance.