Arizona Diamondbacks Atlanta Braves Baltimore Orioles Boston Red Sox Chicago Cubs Chicago White Sox Cincinnati Reds Cleveland Indians Colorado Rockies Detroit Tigers Houston Astros Kansas City Royals Los Angeles Angels Los Angeles Dodgers Miami Marlins Milwaukee Brewers Minnesota Twins New York Mets New York Yankees Oakland Athletics Philadelphia Phillies Pittsburgh Pirates San Diego Padres San Francisco Giants Seattle Mariners St. Louis Cardinals Tampa Bay Rays Texas Rangers Toronto Blue Jays Washington Nationals
 << Previous Article Best of BP 2016: The F... (12/30) << Previous Column Best of BP 2016: The F... (12/30) Next Column >> Best of BP 2016: Let B... (03/13) Next Article >> Fantasy Content Previe... (01/03)

December 30, 2016

# Best of BP 2016

## What We Know About Spin Rate

With the year winding to a close, Baseball Prospectus is revisiting some of our favorite articles of the year. This was originally published on September 27, 2016.

Now that Statcast has begun making data from its system publicly available, baseball fans and analysts alike have a treasure trove of information to look at when analyzing players. The inherent problem that comes from this, of course, is that we don’t know what we don’t know. In order to better understand what exactly this new data is telling us we need to take a look at what we do and don’t know about it.

We’re going to do just that for spin rate, with the goal of clarifying what we know for sure, what we think we know, and what we don’t know or might want explore next.

## Spin Rate 101

What is spin rate?

Spin rate is a measurement (an observed quantity if you want to be technical). It’s an observation of how many times a ball rotates en route to home plate. I would recommend that we not refer to spin rate as a statistic/stat, because it’s simply a measurement. Ewan Ross of BP Toronto noted that we should think about spin rate like we do home run distance, which is a great way to think about it.

How is spin rate measured?

There are actually two methods for measuring spin rate that produce different results.

Method 1 (PITCHf/x) – We are able to calculate spin rate based on the trajectory of each pitch, which is exactly what the PITCHf/x system captures. If we know the pitch’s velocity, starting point (release point), and ending point (where it crosses home plate), then we can calculate how many times it should have spun.

Method 2 (Statcast) – Spin is observed through Trackman’s radar system, with the number of times a ball spins en route to home plate captured and tracked.

What is the difference between calculated and observed spin?

Observed spin should (theoretically) always be higher than the calculated spin. This is because each pitch has a combination of useful spin—spin with a spin axis that is perpendicular to the direction of motion—and gyro spin—spin with a spin axis aligned with the direction of motion. Useful spin results in movement and is isolated by calculating the spin rate from the observed movement of the pitch. Total spin (useful spin plus gyro spin) is captured by Statcast and its observed spin numbers. It is the Pythagorean sum of useful and gyro spin, meaning that it must be greater than or equal to either calculated or observed spin.

What does it mean if observed spin is higher?

It means that having the highest spin rate doesn’t necessarily mean your pitch will move the most. If enough of that spin is aligned with the direction of motion, then spin rate would be misleading in identifying the quality of the pitch.

How should we think about spin?

We should think about spin the same way we think about pitch movement. When we talk about how a pitch moves we inherently have a sense of how pitch type impacts that movement. For example, we know that a four-seam fastball has a different movement profile compared to a slider. We should think of spin the same way. Fastball spin is different from breaking ball spin for a variety of reasons. All spin is not alike.

How much do pitches actually spin?

Probably a lot less than you think they do. Look at this GIF of an Oliver Drake forkball:

Drake’s pitch spins between seven or eight times on the way the home plate.

 Pitch Speed Flight Time Rotations @ 2000 rpm Rotations @ 2500 rpm 100 mph .413 seconds 13.8 17.2 90 mph .458 seconds 15.3 19.1 80 mph .516 seconds 17.2 21.5 70 mph .589 seconds 19.6 24.6

This table helps highlight a different way to think about the spin rate data provided by systems like PITCHf/x or Statcast. The rpm model of presenting the measurements is based on making the data familiar and easy to process. That said, “minutes” is a bad way to measure a pitch whose flight time is roughly half a second. The number of times a ball actually rotates depends on its flight time (velocity) and spin rate. The table above helps illustrate that relationship.

The rightmost column is a helpful guide to consider when looking at spin rate on multiple pitches. If two pitches come in at the same velocity, but differ in spin rate by a few hundred rpm, we can now more easily understand what that means in reality. To make the example more concrete we can look at the following example:

Pitch 1 – 80 mph, 2400 rpm = 20.6 rotations

Pitch 2 – 80 mph, 2100 rpm = 18.1 rotations

You can use this rough logic to back into the spin rate for any pitch if you know some of the data. For example, the Oliver Drake pitch above traveled at 84.1 mph and spun, let’s say eight times. That means the ball was in the air for 0.49 seconds, and spun at roughly 975 rpm.

*****

## Spin Rate 201

Now that we know the basics it’s time to dig deeper into the physics at play. Spin rate is a measurement, as we noted earlier, so it’s important that we understand what and how it’s measured. On top of that, there are other facets of each pitch that directly impact the effect of the spin rate. Each component is worthy of a more thorough examination, which is what we’ll attempt to do below:

The Knuckleball Enigma

Knuckleballs are a fascinating phenomenon that flies in the face of traditional thought around pitch movement and flight. Their secretive nature is further enhanced by the fact that good knuckleballs can’t properly be tracked by a system like Trackman. In fact, pitches that spin at fewer than 500 rpms often aren’t properly tracked within the system.

Likewise, the PITCHf/x values for their spin rate is wrong as well. PITCHf/x—which as we established earlier identifies spin by working backwards from movement—does not know how to handle the fact that a knuckleball can break without spinning. Thus we get incorrect data.

If you’re interested in how knuckleballs move and what PITCHf/x or Trackman can tell us about them, I’d recommend a masters course in the subject from Dr. Alan Nathan. Among the topics covered by Dr. Nathan on his site are:

Other Studies Using PITCHf/x

And of course my personal favorite:

How much spin is “useful” i.e., generates movement?

This varies by pitch and by pitcher. For example, among four-seam fastballs we can see wildly different ratios. Bartolo Colon’s fastball has 60 percent useful spin while Gerrit Cole’s four-seam fastball has 78 percent useful spin. That said, in general fastballs have the highest ratio of useful spin:total spin and sliders have the lowest ratio. This means that sliders have a lot of spin that doesn’t do anything for the movement of the pitch, whereas fastballs do not.

A table below outlines the specific ratios by pitch type, along with the minimum and maximum pitchers (min. 100 pitches*, rounded to nearest 5 rpm):

 Pitch Avg. Total Spin Avg. Useful Spin Useful:Total Spin Min Pitcher, Ratio Max Pitcher, Ratio Four-seam 2,265 rpm 1,605 rpm 71% Robert Whalen, 28% 12 Tied, 100%** Sinker 2,135 1,585 75% Luke Gregerson, 26% 16 Tied, 100%** Cutter 2,300 610 27% A.J. Morris, 8% Josh Fields, 78% Changeup 1,755 1,300 74% Matt Andriese, 13% 4 Tied, 100%** Curveball 2,475 1,490 60% Derek Holland, 8% 3 Tied, 100%** Slider 2,350 665 28% Daniel Coulombe, 6% 2 Tied, 67% Splitter 1,465 830 57% Mat Latos, 6% 2 Tied, 75%

*The pitches excluded from this table are slow curves and knuckleballs (only one pitcher has met our sample size threshold this season for each pitch.)

**Occasionally data issues result in percentages over 100 (meaning the useful spin rate was greater than the total spin rate). This is not technically possible, and is likely an issue in data capture.

Spin axis, the missing piece

Spin axis is critically important when thinking about and discussing spin rate. The reason for this is because it is one of the most critical factors impacting the effect of the spin on the ball. First, a quick overview of the forces acting on the ball, courtesy of Alan Nathan:

The ball is pulled to the ground by gravity (mg) while traveling along a path to the plate identified by the velocity vector (v). The ball is also acted upon by the force of lift (Flift) and the force of drag (Fdrag) as it passes through the air. Finally, the angular velocity vector (w) is determined by the spin axis and spin rate, and it can counteract any/all of the other forces acting on the ball.

The spin axis can impact a pitch greatly. Take a look at the spin axis differential on these two fastballs:

Image courtesy of Sons of Sam Horn (@Soshbaseball)

The slightly tweaked spin axis (indicated by the dotted red line) on the two-seam fastball (aka sinker) is what gives that pitch more arm-side run compared to a four-seam fastball. Sinkers also “drop” more than a four-seam fastball because of a generally lower spin rate, which results in the ball fighting the pull of gravity less than a higher-spin pitch would.

Of course, all this assumes that the spin axis is stable throughout flight, which isn’t always the case. It is possible to identify the spin axis with the naked eye, though you’re largely noticing one of the "poles" of the ball through which it is spinning. This is what creates the “red dot” on a slider, where the hitter sees a red spot on the ball when the spin axis pierces the laces on the underside of the ball.

The video below is a terrific slow motion video that lets you practice at picking up the spin axis around which the ball rotates.

This video showcases Maggie Gallagher, a University of Washington softball player who happens to spin a mean curveball. You can see that the spin axis on her breaking ball is incredibly stable from release throughout its flight path, making it a perfect case study for someone learning and identifying spin axes in the real world.

The big takeaway is that spin rate works in conjunction with spin axis. Spin perpendicular to the direction of motion generates movement (useful spin), but other spin (gyro spin) doesn’t necessarily help the pitch in terms of movement. So sure, try to get that spin rate up there; if it’s not in alignment with the spin axis though, it’s not doing anything for the movement of that pitch.

How much does spin rate matter?

In order to answer this question, I decided to simply look at correlations between spin rate and a variety of outcomes for each type of pitch. The first table includes the results for total spin, while the second table includes only useful spin.

 Total Spin (Statcast) Correl: SwStr% Correl: Called Strike Prob. Correl: GB% Correl: FB% Four-seam -0.06 0.10 -0.05 0.03 Sinker -0.06 0.11 -0.05 0.03 Cutter -0.06 0.10 -0.05 0.03 Changeup -0.06 0.10 -0.05 0.03 Curveball -0.06 0.10 -0.06 0.03 Slider -0.07 0.10 -0.06 0.04 Splitter -0.07 0.11 -0.06 0.04

 Useful Spin (PITCHf/x) Correl: SwStr% Correl: Called Strike Prob. Correl: GB% Correl: FB% Four-seam -0.09 0.10 -0.02 0.02 Sinker -0.10 0.11 -0.02 0.02 Cutter -0.10 0.10 -0.01 0.01 Changeup -0.09 0.10 -0.02 0.02 Curveball -0.10 0.10 -0.01 0.01 Slider -0.09 0.12 -0.02 0.01 Splitter -0.09 0.13 -0.02 0.00

The results here are pretty clear. There’s no strong relationship between spin rate and any of these outcomes. Also, you could argue that total spin isn’t an improvement on useful spin data (which has been calculable for years) given that the correlations are fairly similar across both tables. This of course shouldn’t be too much of a shock because, while spin rate is a great measurement to have, ultimately pitch movement and velocity impact the outcome more than how many times the ball spins on its way to the catcher.

There is an interesting nugget in here, though. You’ll notice that the correlations are fairly similar across each pitch type, which suggests that spin rate might be an innate characteristic. That is, high-spin pitchers on their fastball also tend to be high-spin pitchers for their breaking balls, changeups, etc. The fact that the correlation for these various outcomes is nearly static across various pitches and various average spin rates (as we established previously) is fascinating. Over the population of MLB pitchers, spinning the ball seems to be something some pitchers excel at while others falter.

This is really only enough to suggest that this might be worthy of study, as there isn’t enough evidence here to definitively proclaim that spinning the ball is an innate skill.

Fastball spin and velocity

Kyle Boddy of Driveline Baseball has done some of the most advanced work on spin rate with the assistance of Trevor Bauer of the Cleveland Indians. Boddy and Bauer have coined the term “Bauer Units,” used to describe the relationship between fastball velocity and spin rate. Simply put, Bauer Units can be calculated as “Pitch Spin (rpm) / Pitch Velocity (mph).”

Boddy has worked with the pitchers who train at his facility to further study spin rate and velocity, as well as their relationship. In a limited study (12 pitchers) Boddy found that the relationship is roughly linear. An increase in velocity resulted in an equal increase in spin rate. Boddy discusses their research in much more depth on the Statcast podcast, which can be found here.

Boddy would be the first to admit, however, that this study is limited in scope, and as such is merely a starting point. Ideally it would be replicated by Boddy as well as others with more pitchers. Ideally it would take other aspects into account (which Boddy may have done, though he hasn’t acknowledged so publicly) like pitch movement, spin ratios, etc.

Kyle Boddy and Trevor Bauer know more about fastball spin rate than just about anyone out there (not in a front office, anyway) and yet they’re still scratching the surface of what we can learn.

## Spin Rate 301

This is a convenient segue into the next section of spin-rate analysis. This is the section where we highlight what we don’t know about spin rate. This is where we take a look at all the things that we are interested in, but haven’t yet discovered or proved.

I sent out a call on Twitter for what we wanted to know about spin rate. Here are some of the requests for which we don’t have answers:

These are just the tip of the iceberg so if you’re interested in spin rate and you’ve made it this far, go grab some data, and test some of these hypotheses/questions. We need more informed analysis of spin rates, so let’s continue to build on the knowledge we already have!

PS - For those interested in the ties for Useful:Total Spin ratio, they are:

Four-seam (12 @ 100%): Sean Manaea, Jorge De La Rosa, Robbie Ray, Pat Dean, Hector Santiago, Chris Sale, Grant Dayton, Danny Duffy, Matthew Strahm, Jake McGee, Antonia Bastardo, Tyler Anderson

Sinker (16 @ 100%): Matt Purke, Richard Bleier, Adam Morgan, Wade LeBlanc, Mike Montgomery, Alexander Claudio, Justin Wilson, Aaron Loup, Martin Perez, David Price, Chris Sale, Danny Duffy, Oliver Perez, Ross Detwiler, Jake Diekman, Robbie Ray

Changeup (4 @ 100%): J.A. Happ, Ariel Miranda, Adam Conley, Chris Sale

Curveball (3 @ 100%): Luke Hochevar, Adam Wainwright, Chris Tillman

Slider (67%): Corey Kluber, Andrew Triggs

Splitter (75%): Mike Pelfrey, Hector Neris

Special thanks to Alan Nathan, Harry Pavlidis, Rob McQuown, Kate Morrison, Sean O'Rourke, Neil Weinberg, Evan Davis, Daren Willman, and many others for their assistance in compiling the information for this post.

Jeff Long is an author of Baseball Prospectus.
Click here to see Jeff's other articles. You can contact Jeff by clicking here

Related Content:  Spin Rate

 << Previous Article Best of BP 2016: The F... (12/30) << Previous Column Best of BP 2016: The F... (12/30) Next Column >> Best of BP 2016: Let B... (03/13) Next Article >> Fantasy Content Previe... (01/03)

 RECENTLY AT BASEBALL PROSPECTUS Rubbing Mud: Lineups Are Just a Social Const... Fantasy Starting Pitcher Planner: Week 20 Weekly Wrap: August 18, 2017 Banjo Hitter: The First 162: Alex Bregman Guarding The Lines: Here The Fastball Are No... Cold Takes: Doomed and Determined Circle Change: The Best Failure in Baseball MORE FROM DECEMBER 30, 2016 Best of BP 2016: The Fall of the Ryan Empire Best of BP 2016: Making a Pitch for Practice... Best of BP 2016: Hardware MORE BY JEFF LONG 2017-01-25 - Prospectus Feature: Two Ways to Tunnel 2017-01-24 - Prospectus Feature: Introducing Pitch Tunnel... 2017-01-23 - Prospectus Feature: Command and Control 2016-12-30 - Best of BP 2016: What We Know About Spin Rat... 2016-11-25 - Pitching Backward: Reviewing Rapsodo 2016-10-05 - Pitching Backward: Closing the Window 2016-09-27 - Pitching Backward: What We Know About Spin R... More... MORE BEST OF BP 2016 2017-03-13 - Best of BP 2016: Let Ballparks Get Old 2016-12-30 - Best of BP 2016: What We Know About Spin Rat... 2016-12-30 - Best of BP 2016: The Fall of the Ryan Empire 2016-12-30 - Best of BP 2016: Making a Pitch for Practice... 2016-12-30 - Best of BP 2016: Hardware More...