Tactics
Pitch design in baseball explained: engineering movement
Pitch design in baseball explained: what trackers measure, why spin efficiency beats raw spin, how seam-shifted wake broke the old model.
By CricketTaken EditorialPublished Tactics22 min read
Twenty years ago a pitcher found out what his curveball did by throwing it and watching. If it broke, he threw it more. If it hung, a coach told him to get on top of it, which is advice that means almost nothing and was passed down for a century anyway. Stuff was something you were born with, discovered in your teens and then either kept or lost.
That is over. Pitch design in baseball is now a manufacturing process, and the phrase covers a real workflow: measure what the ball is doing, decide what it ought to be doing, change the grip and the release until the measurement moves. A pitcher can arrive at a facility in October with four pitches and leave in January with three of them rebuilt and a fifth that did not previously exist.
The change did not come from better coaching. It came from being able to see the ball. Once a camera could report the direction a baseball was spinning, in three dimensions, sixty times a second, the whole conversation stopped being about feel and started being about physics.
What follows is the mechanism underneath that, including the part where the model everyone had just learned turned out to be incomplete.
What the trackers actually measure, and what they infer
Every pitch in a professional stadium is now tracked by a camera array rather than a radar gun, and the distinction matters because it changed which numbers are measured and which are calculated.
The system reports the ball's position many times between release and the plate, and from that it derives the trajectory. Release speed. Release point in three dimensions. Extension, meaning how far in front of the rubber the ball is actually let go. Total movement, split into horizontal and vertical components in inches. Spin rate in revolutions per minute.
Then there is the number that reorganised the field: spin axis, measured directly rather than worked out backwards.
- 60.5Feet from the pitcher's plate to home plate
- 17Width of home plate, in inches
- 8.5Depth of the automated strike zone, in inches from front and back
- 2Ball and strike challenges each team gets per game
All four are set by rule rather than by convention, and none of them change from season to season.
Older systems could not see the seams well enough to report an axis, so they inferred it from the movement: the ball ended up here rather than there, therefore it must have been spinning that way. That inference is only valid if spin is the sole cause of movement.
Camera systems can watch the ball turn. They report a measured axis, expressed as a position on a clock face, where a four-seam fastball thrown by a right-hander with backspin sits near one o'clock and a pitch with pure sidespin sits at three or nine. They also still report the movement-inferred axis, calculated the old way.
Two numbers for the same thing. When they disagree, something other than spin is moving the ball, and the size of the disagreement is the single most useful diagnostic in modern pitch design. That is not an accident of the data. It is the reason the data is published in that form.
Spin rate is the famous number and spin axis is the useful one
Spin rate got the attention because it is a single figure that sounds like a talent. High spin, good pitcher. It has been on broadcast graphics for years and it is the number a pitcher's agent will quote.
It is close to useless on its own.
Spin rate tells you how much rotational energy the ball is carrying, and therefore how much force is potentially available. It says nothing about the direction of that force, and direction is what makes a pitch a pitch. A fastball that resists gravity and a slider that dives are separated by axis, not by rate. Raw spin rate on its own is an ingredient, not a description.
Axis is the direction the ball's rotation points, and the force it produces, the Magnus force, acts perpendicular to both the axis and the direction of travel. Backspin pushes the ball up relative to where gravity would take it, which is the "ride" on a four-seam fastball. Topspin pushes it down, which is a curveball. Sidespin pushes it sideways.
Coaches think in clock faces because the notation is legible and because two pitchers with different arm slots can compare notes in the same units. A right-hander who wants more carry on his fastball is being asked to move his axis closer to twelve o'clock, which is a physical instruction about what his hand does, not a metaphor about getting on top of the ball.
There is a second reason axis dominates the conversation. Rate is very hard to change. A pitcher's spin rate on a given pitch type is largely a product of grip strength, finger friction, arm speed and hand size, and years of training move it a little. Axis can be changed in an afternoon by moving a finger.
If you are engineering a pitch, you work on the lever that moves.
Spin efficiency, and the spin that does nothing at all
Here is the concept that separates people who have read about pitch design from people who do it.
Not all spin produces movement. The Magnus force depends on the component of the spin that is perpendicular to the direction the ball is travelling. Any component pointing along the direction of travel produces no force whatsoever. It turns the ball like a rifle bullet, stabilising it and doing nothing else. That is gyroscopic spin, and pitchers call it gyro.
Spin efficiency, published in the public data as active spin, is the fraction of total spin that is doing the useful perpendicular work. A pitch at 100 per cent efficiency has its axis square to its path and gets every ounce of movement its rate can produce. A pitch at zero is a perfect bullet spin, spinning fast and moving nowhere.
This is why a fastball at 2,600 rpm can have less ride than one at 2,200. If the faster-spinning ball is throwing away a third of that rotation into gyro, the slower one with a cleaner axis wins.
- Transverse spin, producing Magnus force1870
- Gyroscopic spin, producing no force at all330
Invented round numbers, used to show the decomposition. Efficiency is defined as the share of total spin acting perpendicular to the ball's direction of travel; the remainder is gyroscopic and produces no Magnus force.
Show the numbers
| Item | Value |
|---|---|
| Transverse spin, producing Magnus force | 1870 |
| Gyroscopic spin, producing no force at all | 330 |
Sliders are where this stops being an abstraction. A traditional slider is thrown with a great deal of gyro, sometimes most of its spin, which is why it does not move very much through the air and why its bite is so late: it holds a fastball line and then simply stops resisting gravity in the way a fastball does. A sweeper is the opposite design, the same family of pitch retuned so that far more of the spin is transverse and pointed to produce horizontal break.
Both are called sliders. They are different objects, and the number that separates them is efficiency, not rate.
The practical upshot for a coach is that "throw it harder" and "spin it more" are usually not available instructions, while "spin it in a different direction" and "waste less of it" often are. Efficiency is the middle lever, and it responds to grip changes.
Seam-shifted wake, the effect that broke the model
Everything above was the settled understanding by the late 2010s, and then a set of pitches turned out not to fit it.
Certain pitchers, throwing sinkers and changeups with unremarkable spin rates and moderate efficiency, were getting substantially more arm-side movement than the Magnus force could account for. The measured axis said one thing. The movement said another. The gap was too large and too repeatable to be noise.
The explanation came out of a university aerodynamics lab rather than out of baseball, and it is worth stating precisely because the popular version of it is wrong.
A baseball flying through the air carries a thin layer of air dragged along with its surface, the boundary layer. That layer separates from the ball somewhere near the back, and where it separates determines the shape of the wake behind the ball. A wake that is symmetrical produces no sideways force. A wake that is pushed off to one side produces a force in the opposite direction, in the same way that a fire hose pushes a firefighter backwards.
A raised seam sitting in the right place disturbs the boundary layer and makes it separate earlier on that side of the ball. The wake shifts towards the seam. The ball is pushed away from it.
That last sentence is the part people get backwards, and it is why the earlier explanations borrowed from cricket were abandoned. The first attempts to describe the effect reached for laminar and turbulent flow, the mechanism behind swing bowling in cricket, which produces a deflection towards the disturbed side. Baseball's seams do the opposite. The dominant effect is the removal of the boundary layer on the seam's side, and the deflection runs away from the seam. Similar phenomenon, opposite sign, and the distinction matters because it flips which grip you would reach for.
Now the part that ties it to gyro spin, which is what makes the whole thing usable.
For a seam effect to accumulate over sixty feet, the seam has to stay roughly in the same place relative to the airflow. On a highly efficient four-seam fastball it does not: the ball is rotating square to its path, so any seam that is disturbing the flow on one half of a revolution is disturbing it on the other side half a revolution later, and the forces cancel. On a pitch carrying significant gyro spin, the ball is rotating about an axis close to its direction of travel, and a seam can be held pointing the same way for the entire flight.
Low efficiency, which the old model treated as pure waste, turned out to be the precondition for a second source of movement. That is why the field's attitude to gyro flipped from "spin that does nothing" to "spin that does nothing by itself".
Reading movement against the spin-only baseline
The measurement that makes seam effects visible is a comparison, and it is available publicly.
Take a pitch's measured spin axis and its spin rate, and calculate the movement it should have if Magnus were the only force acting on it. Then compare that to the movement the tracking system actually recorded. The difference is everything spin cannot explain, and in practice that means seam effects.
The fastball in that example is behaving exactly as the model says. The sinker and the changeup are not, and the four inches of unexplained arm-side run on each is what a coach is hunting for. Four inches is roughly half a ball width at the plate, which is the difference between contact on the barrel and contact on the handle.
Two things follow from being able to make that comparison.
The first is diagnosis. A pitcher whose sinker has stopped working can now be told whether the problem is spin, in which case the axis has drifted, or seams, in which case the ball is coming out of the hand oriented differently even though the spin looks unchanged. Those two failures feel identical to the pitcher and require opposite fixes.
The second is design. If you know a pitch is getting four inches from its seams, you know that orientation at release is a variable worth training, and orientation at release is coachable in a way that spin rate is not. Practice devices now report seam orientation directly, which was not commercially available a few years ago.
The three levers: grip, wrist and the last inch
All of this is theory until someone has to throw differently. There are only three places a pitcher can intervene, and they do different jobs.
Grip changes the axis and the seam orientation. Moving a finger from on top of a seam to alongside it, rotating the ball a quarter turn in the hand, spreading or narrowing the two fingers: these change where the pressure is applied and therefore which way the ball is turning when it leaves. Grip is the cheapest change and the first one tried, because it can be altered between pitches and reversed instantly if the number goes the wrong way.
Wrist and forearm position change efficiency. Whether the hand is behind the ball, beside it or turned over at release determines how much of the spin ends up transverse and how much ends up gyro. This is a harder change because it involves the arm action rather than the hand, and an arm action that has been grooved for a decade does not move without repetition. It is also the change most likely to cost velocity in the short term.
Release point and extension change the perceived version of everything. A ball let go six inches further down the mound arrives sooner at the same speed, which raises the perceived velocity, and it changes the angle the pitch enters the zone at. Two pitchers with identical stuff and different arm slots present completely different problems, because the vertical approach angle into the strike zone is a function of where the ball started.
The order matters. A good session tries grips first, because they are reversible and fast, and only moves to arm action when the grip search has been exhausted. Changing the arm action to get two inches of break, and losing a mile an hour and a month of command doing it, is a bad trade that gets made more often than anyone admits.
One recent grip family shows how quickly this now moves. A changeup thrown with a deliberate kick from one finger, producing gyro spin and a specific seam orientation rather than the traditional pronated fade, went from a curiosity to a widely taught pitch in a couple of years. It spread because the mechanism was legible: coaches could see on a screen what it was doing and why, and they could teach it from the numbers instead of from a feel that only one pitcher had.
An arsenal is designed as a set, not as five separate pitches
The single most common error in amateur pitch design is optimising each pitch on its own. A pitcher ends up with five individually excellent pitches that a hitter can tell apart at release, which is worse than three that he cannot.
The organising idea is the tunnel. A hitter cannot wait for the ball to arrive; he has to decide to swing while the pitch is still a long way out, and once he has committed, he is guessing. Everything that happens after his decision point is free movement as far as the pitcher is concerned. Everything before it is information.
So a well-designed pair of pitches shares a trajectory for as long as possible and then separates. A fastball and a slider that come out of the hand at the same angle and stay within a ball's width of each other for the first half of the flight are a genuine problem, because the hitter is choosing between them on a difference he cannot see. The same fastball and a slider released from a visibly different slot are two pitches, not a pair.
Pairing works on the same principle at a larger scale. A pitcher wants his pitches to cover different quadrants of the plate from a common start, so that the hitter's guess is punished in more than one direction. The classic construction is a fastball at the top of the zone and something with the opposite vertical movement beneath it, with a third pitch running the other way horizontally.
This is where the strike zone becomes a design constraint rather than a backdrop. The zone as the rules define it is a volume, and the movement profile that makes a pitch unhittable at the top of it is not the one that works at the bottom. A four-seamer with heavy ride is a weapon at the letters and a batting practice fastball at the knees, because at the knees the ride carries it into the middle of the zone.
There is a scheduling element too. Pitches are not independent events, and a hitter who has seen four fastballs is timing a fastball. The value of a designed pitch is partly the pitch and partly its position in the sequence, which is why arsenals are built with an eye to what the pitcher will actually throw in a three-two count against a left-hander rather than to what looks best on a spreadsheet.
- Baseline captureThe pitcher throws his existing arsenal into a tracking unit and in front of a high-speed camera. Every pitch is logged with velocity, spin rate, measured spin axis, spin efficiency, release point, extension and movement in both planes.
- DiagnosisThe measured movement is compared against what the spin alone predicts. A pitch that matches the prediction is a pure Magnus pitch. A pitch that beats it is getting help from its seams, and one that falls short is losing something at release.
- Choosing the targetThe coach decides what the arsenal is missing rather than what any single pitch could be. Usually that is a movement direction nobody currently covers, or a pitch that leaves the hand looking like an existing one.
- Grip searchSeveral grips are tried in short sets, each one measured. The pitcher is not asked how it felt until after the numbers are read, because feel is the least reliable instrument in the room and it lags the change by weeks.
- Video checkHigh-speed footage at hundreds of frames per second shows the hand at release, the seam orientation and whether the ball is being cut or pushed. This is the step that explains why a grip failed.
- Constraint checkThe candidate pitch is judged against the rest of the arsenal, not on its own. Does it share a tunnel with the fastball? Can it be thrown for a strike? Does it cost velocity elsewhere?
- Repetition and commandA shape that works once is a party trick. The pitch is thrown in volume until the movement is repeatable and it can be located, which is the stage most designed pitches die at.
- Transfer to competitionThe pitch is used against live hitters, then in games, and the tracking data is checked again. Shapes routinely degrade under game intensity, and the version that survives is usually a compromise with the one designed in the lab.
The workflow used in a lab or a team facility. The order is the point: measurement, target, intervention, verification, then transfer to a game.
The hitters have the same screens
Pitch design is presented as a pitching revolution, which is half the story. Hitters have the same measurement, the same laboratories and, increasingly, the same coaches.
Bat tracking data is now public in the same way pitch tracking is. Swing speed, swing length and how squarely the ball was struck are all measured, which means a hitter's swing can be engineered against a specific pitch shape with the same feedback loop a pitcher uses. A hitter who cannot cover the top of the zone against ride now knows it as a number rather than as a suspicion, and he can work on it deliberately. The batted-ball side of the same measurement, how hard and at what angle the ball leaves the bat, is what tells him whether the change worked.
Advance scouting closed the other gap. Every pitcher's movement profile, release point and sequencing tendencies are available before he throws a pitch, so the element of surprise that a new pitch used to carry has a shelf life measured in starts. Design a sweeper in January and by June the league has seen it, charted it and started spitting on it.
This produces the pattern you would expect from an arms race: the advantage is real and it is temporary. A designed pitch is worth the most in the season it appears and less in every one after, which is why the facilities exist and why pitchers go back to them every winter. The equilibrium is not "pitchers won". It is "both sides now iterate faster, and neither can stop".
The automated strike zone adds a new constraint on the pitcher's side of that race. Under a challenge system, the definition of a strike stops being partly a negotiation with the umpire and becomes, for the pitches that get challenged, a geometric fact about where the ball crossed a plane at the midpoint of the plate. That changes what a pitch on the black is worth. The way catchers earn strikes at the edges has been one of the quiet inputs into pitch design for a decade, because a shape that a catcher can receive cleanly is worth more than one he has to reach for. Handing part of that judgement to a tracking system does not remove the incentive, but it does put a hard boundary on how far outside the zone a designed breaking ball can finish and still be worth throwing.
The injury argument, characterised honestly
This is the section a piece on pitch design cannot skip, and it is also the one where it is easiest to overstate the evidence in either direction.
The facts that are not in dispute: pitchers throw harder than they used to, the increase has been steady over two decades, and serious elbow injuries in professional pitchers have risen over the same period. The elbow's ulnar collateral ligament is loaded by the valgus torque generated as the arm lays back and accelerates, and that torque scales with how fast the arm is moving. Faster arm, more torque, closer to the ligament's failure threshold on every pitch.
Major League Baseball ran a lengthy review of the problem and published its conclusions. It drew on a large body of expert input, including surgeons, biomechanists, pitching coaches and former players, and its judgement was that the most likely causes were the rise in velocity, the pursuit of optimised movement profiles, and the modern habit of throwing at maximum effort in games and in training alike. The report recommended that the league consider rule changes that would increase the value of durability and reduce the reward for short, maximum-effort outings.
Separately, published work using the public tracking data has reported that pitchers who went on to have ligament reconstruction had, on average, thrown harder and with more spin than those who did not.
Now the honest characterisation, which matters more than the headline.
None of this is experimental. Nobody has randomised a group of pitchers into a high-effort arm and a low-effort arm, and nobody is going to. The evidence is observational, and observational data on this question has an obvious confound running through it: the pitchers who throw hardest are also the pitchers who get promoted, pitch the most innings, face the best hitters and are scrutinised most closely when their elbow hurts. Velocity is correlated with opportunity, and opportunity is correlated with exposure.
There is a second confound in the training pipeline. The same period saw a large increase in year-round youth baseball, radar guns at high school showcases, and velocity programmes for teenagers whose growth plates have not closed. A pitcher who reaches the major leagues today has usually thrown more maximum-effort pitches by the age of twenty than a pitcher of an earlier generation had by twenty-five. Whether the damage is done in the professional bullpen or in the decade before it is a genuinely open question, and the answer determines which rule changes would help.
Where does pitch design itself sit in that? Its specific contribution is hard to isolate, because the pursuit of movement and the pursuit of velocity travel together. A pitcher chasing a higher spin rate is generally also gripping harder and throwing harder. Some designed pitches, particularly those that rely on pronounced supination, have been suspected of loading the elbow differently, and that suspicion is plausible and not established.
The defensible summary is this. The association between maximum-effort pitching and elbow injury is strong, consistent and mechanistically sensible. The league itself has said so and has said it publicly. The share of that attributable to pitch design as distinct from velocity, workload and youth exposure is not known, and anyone who tells you they can put a number on it is telling you something the data does not support.
That uncertainty cuts both ways, and it is why the interesting policy proposals target incentives rather than techniques. You cannot regulate a grip. You can change what a team gets for using a pitcher in one-inning bursts at full effort, and rules that reward starters going deeper would do more to the training culture than any ban on a pitch type.
Watching a start with this in mind
The next time you watch a pitcher work, four things are visible from the couch once you know to look for them.
The release point of every pitch he throws. If the slider comes from a slightly lower slot than the fastball, hitters know it before the ball is halfway. A pitcher whose whole arsenal leaves from one window is a designed pitcher, whatever the movement numbers say.
Whether the fastball is being used at the top of the zone or the bottom. A four-seamer with ride is thrown up on purpose, and a sinker is thrown down on purpose. A pitcher throwing the wrong one to the wrong height is either missing his spots or has been given a plan that does not match his shapes.
Which two pitches he pairs in two-strike counts. That is his tunnel, and it is the pair he trusts. The other pitches are there to set it up or to steal a strike early.
Whether a new pitch appears in June that was not there in April. That is a design job landing mid-season, and it is now common enough that a scouting report from spring training goes stale.
And one number, if you want to check the work: pull up the pitcher's page on the public tracking site and look at his spin direction, where the measured axis and the movement-inferred axis sit side by side. A large gap between them means he is getting movement his spin cannot explain, which means his seams are doing work. That is not a curiosity. It is the single most direct evidence that somebody, somewhere, sat him down in front of a camera and built the pitch on purpose. The rest of the tactical and analytical explainers here start from the same place: find the measurement, then ask what it made people do differently.
Common questions
What is pitch design in baseball?
Pitch design is the deliberate engineering of a pitch's shape using measured data rather than feel. A pitcher throws into a tracking system that reports velocity, spin rate, spin axis, release point and movement, a coach compares that against what the pitch would need to look like to work, and grip, wrist position and release are then adjusted until the numbers move. It replaced an older model in which a pitcher essentially inherited whatever his arm produced and learned to live with it.
What is the difference between spin rate and spin axis?
Spin rate is how fast the ball is turning, measured in revolutions per minute, and spin axis is the direction that rotation points, usually described as a position on a clock face. Rate tells you how much force is available and axis tells you which way that force pushes the ball. Two pitches with identical spin rates can move in completely different directions, which is why axis is the more useful of the two numbers for shaping a pitch.
Why does spin efficiency matter more than spin rate?
Only the component of spin that is perpendicular to the ball's direction of travel produces a Magnus force, so a ball can be spinning very fast and still barely deviate. Spin efficiency, also published as active spin, is the fraction of total spin doing that useful work, and the rest is gyroscopic spin that turns the ball like a rifle bullet without pushing it anywhere. A slower-spinning pitch with high efficiency will out-move a faster-spinning one with low efficiency.
What is seam-shifted wake?
Seam-shifted wake is movement produced by the seams rather than by spin. When a seam sits in the right place relative to the airflow it pulls the boundary layer off the ball earlier on one side than the other, and that asymmetric separation pushes the ball away from the seam. Because the effect depends on where the seams are pointed, a pitch with a lot of gyroscopic spin can hold a seam in the same orientation for the whole flight and accumulate a force that spin alone cannot explain.
Does pitch design cause pitcher injuries?
Major League Baseball's own review of the rise in pitcher injuries concluded that increased velocity, the pursuit of optimised stuff and the habit of throwing at maximum effort in games and in training are the most likely drivers, and it recommended that rule changes be considered to reduce the reward for short, maximum-effort outings. That is a considered judgement from a large body of expert opinion rather than a controlled experiment, and the underlying data is observational, so the honest position is that the association is strong and consistent while the causal share attributable to pitch design specifically is not established.
Filed under Baseball·baseball · mlb · pitching · analytics · statcast · coaching