Spin rate in baseball explained: rpm is only half of it

Two pitchers throw a fastball. Both are tracked at the same revolutions per minute. One of the pitches finishes at the top of the zone having barely dropped and gets swung under; the other arrives where the hitter expected it and leaves the park. Nothing in the spin rate figure distinguishes them, and nothing ever will, because the number counts revolutions without asking what the revolutions are doing.

That gap is the whole subject. Spin rate in baseball is genuinely important and genuinely misread, and the misreading has a specific shape: a single rpm figure gets treated as a quality score for a pitch. It is not one. It is one of four numbers that together decide how a pitch moves, and it is the only one of the four that most coverage ever mentions.

What the number is, exactly

MLB's Statcast glossary is unambiguous about the definition. Spin Rate "measures the spin on the baseball when the pitch is released. It is measured in revolutions per minute."

Two words in that sentence carry most of the weight.

Released. The figure is taken at the hand, not averaged across the flight. A ball loses a little rotation on the way to the plate, and none of that is in the number. What you are reading is a snapshot of the instant the fingers came off the seams, which makes it a property of the pitcher's release rather than of the pitch's journey.

Revolutions. It is a count, not a force. Revolutions per minute is a rate of rotation and says nothing about the axis that rotation is happening around, which is the thing that decides which direction the ball gets pushed.

Baseball Savant's own field documentation, which describes every column in the public data, defines the release spin field with the same flatness: "Spin rate of pitch tracked by Statcast." No qualifier, no interpretation. The tracking system reports the rotation; everything else is inference performed downstream.

MLB's glossary also states what the spin does, and it is worth reading closely because it is a description of consequences rather than of causes. The amount of spin on a pitch changes its trajectory, so pitches thrown at identical velocity end up in different places. High-spin fastballs appear to rise several inches higher than low-spin fastballs. High-spin curveballs generate more break. Lower spin rates on changeups and splitters create increased downward tumbling movement.

Read that list again and notice the pattern. In every case the effect depends on what kind of pitch it is. There is no sentence in the definition of the form "more spin is better", because no such sentence would be true.

Why a spinning ball moves at all

The mechanism is old physics and it is worth stating properly, because almost every misunderstanding downstream comes from a fuzzy grasp of it.

A baseball travelling through air drags a thin layer of that air along with its surface. If the ball is also rotating, the surface on one side is moving in the same direction as the airflow and the surface on the other side is moving against it. Air on the with-the-flow side is sped up. Air on the against-the-flow side is slowed down. Faster-moving air exerts less pressure, so a pressure difference develops across the ball and the ball is pushed towards the low-pressure side.

That push is always perpendicular to two things at once: the direction the ball is travelling, and the axis it is spinning about. Fix those two directions and you have fixed the direction of the force. That single geometric fact explains everything that follows.

A ball with pure backspin, rotating so the top of the ball turns back towards the pitcher, gets pushed upwards. It does not rise, because the push is smaller than gravity, but it drops less than a hitter's eye expects from a ball thrown that hard. The eye interprets less drop than expected as rise, which is why the word keeps appearing in quotation marks in official definitions.

A ball with pure topspin gets pushed downwards, and drops more than gravity alone would take it. That is a curveball.

A ball spinning about a vertical axis gets pushed sideways.

Every real pitch is some combination of those three, which is where the axis comes in.

Spin axis, and the number that tells you the direction

Savant publishes the axis as a field of its own, and the definition is precise: spin axis is "the Spin Axis in the 2D X-Z plane in degrees from 0 to 360, such that 180 represents a pure backspin fastball and 0 degrees represents a pure topspin (12-6) curveball."

So the axis is reported as an angle, and two of its landmarks are given to you directly by the definition. Everything between them is interpolation.

• 180°Pure backspin, the four-seam fastball
• 0°Pure topspin, the twelve-to-six curveball
• 90°A quarter turn from either, pure side-spin

Degrees in the two-dimensional plane, as published in Baseball Savant's field documentation. 180 and 0 are given in the definition; the quarter-turn positions are pure side-spin, which follows from the same geometry.

Coaches usually talk about this in clock-face terms instead, describing a pitch as spinning at one o'clock or seven o'clock, because a clock face is easier to hold in the mind and easier to demonstrate with a ball in your hand. The two conventions describe the same thing.

What the axis gives you is direction. What it does not give you is magnitude, and that is where most public discussion of spin quietly runs out.

Spin rate against active spin, which is the distinction that matters

Here is the part that separates a useful reading of spin from a decorative one.

Not all rotation produces force. MLB's Statcast glossary makes the point with an analogy that is better than any explanation: a spiralling American football or a bullet from a rifled barrel is spinning at an enormous rate, and none of that spin deflects it, because the rotation is about the same axis as the direction of travel. That kind of rotation is a hundred per cent inactive.

MLB's definition of the useful quantity is short. Statcast "refers to the spin that contributes to movement as Active Spin." When a ball is thrown with pure topspin or backspin, then a hundred per cent of the spin contributes to movement. A pitcher can also apply side-action, so some of the deflection is vertical and some horizontal, and all of that still counts as active spin.

The complement is gyroscopic spin, rotation about the direction of flight, which contributes nothing. A slider thrown with a lot of gyro spin can be tracked at a high rpm figure and still move very little, because the revolutions are pointing the wrong way.

The ratio of active spin to total spin is what people mean by spin efficiency. It is expressed as a percentage, and it is the single most important companion to the rpm number.

90%10%

• Active spin, producing deflection2160
• Gyroscopic spin, producing none240

Invented figures, chosen round. A pitch tracked at 2,400 rpm with 90 per cent efficiency has 2,160 rpm generating force and 240 rpm doing nothing at all. No real pitcher's spin profile is described here.

Show the numbers

Where a constructed 2,400 rpm actually goes

Item	Value
Active spin, producing deflection	2160
Gyroscopic spin, producing none	240

Now take the two pitchers from the opening paragraph and give them numbers. Both are invented and deliberately round.

Pitcher A is tracked at 2,400 rpm with 95 per cent efficiency. His active spin is 2,280 rpm.

Pitcher B is tracked at exactly the same 2,400 rpm with 60 per cent efficiency. His active spin is 1,440 rpm.

The broadcast graphic shows both of them the same number. The pitches are not remotely the same pitch, and the second one has roughly the deflection budget of a much lower-spin fastball thrown efficiently.

• Total spin
• Active spin

Pitcher A, 95 per cent efficient2400rpm2280rpm

Pitcher B, 60 per cent efficient2400rpm1440rpm

Invented figures. Both are tracked at 2,400 rpm; only the share of that rotation which points somewhere useful differs. These are constructed examples, not measured players.

Show the numbers

Two constructed pitchers, identical spin rate, different pitches

Item	Total spin	Active spin
Pitcher A, 95 per cent efficient	2400rpm	2280rpm
Pitcher B, 60 per cent efficient	2400rpm	1440rpm

This is the reason a pitching coach asks for the efficiency figure before he asks for the rpm figure, and it is the reason a competing page that stops at "high spin is good" is not merely incomplete but occasionally backwards. A high-spin, low-efficiency four-seam fastball is a worse pitch than a moderate-spin, high-efficiency one, and the raw number ranks them the wrong way round.

Where the efficiency figure comes from, and why that matters

Active spin is not read off a sensor the way velocity is. The tracking system reports total spin and the release conditions; the share of it that is active is worked out from what the ball actually did on the way to the plate, compared with what a ball of that speed, spinning on that axis, ought to have done.

That has an awkward consequence, and it is the honest caveat this subject usually goes without.

If the observed movement is the input to the efficiency calculation, then any movement produced by something other than spin will be absorbed into the efficiency figure. Baseballs are not smooth spheres. They have raised seams, and the orientation of those seams relative to the airflow can influence the wake behind the ball. Where that happens, a pitch moves in a way its spin axis alone does not predict, and the inferred axis and the released axis stop agreeing.

That is not a defect in anybody's data. It is a genuine open area, and the useful posture is to treat spin efficiency as an excellent summary of most pitches and an incomplete description of the ones that behave strangely. Pitches with unusual seam orientations are exactly the ones a pitching lab is interested in, and exactly the ones where the public figure understates what is going on. The design of individual pitches as a discipline exists largely to exploit that gap.

How a movement number is actually built

The output of all of this is the movement column, and its construction is more opinionated than it looks.

MLB's glossary states the reporting convention plainly. "The movement of a pitch is defined in inches, both in raw numbers and (more importantly) as a measurement against average." Savant's field documentation gives the raw columns their names: horizontal movement in feet from the catcher's perspective, and vertical movement in feet from the same viewpoint.

Two decisions inside that are worth pulling out.

Gravity is included. MLB says its "pitch movement numbers are displayed with gravity, in an attempt to better align with the real-world eye test." Every pitch therefore shows substantial downward movement, because every pitch falls. A fastball's celebrated rise is a smaller amount of drop, not a positive number.

Average is a matched cohort, not a global mean. This is the detail that almost never gets repeated. Movement is compared with "average" movement, MLB says, "by comparing it to other MLB pitch types thrown within +/- 2 mph of velocity and +/- 0.5 feet of extension and release point."

• 2Velocity band, miles per hour either side
• 0.5Extension band, feet either side
• 0.5Release point band, feet either side

The tolerances MLB publishes for building the comparison group behind a movement-versus-average figure. A pitch is measured against other pitches of its type thrown from a similar place at a similar speed.

That means a movement-versus-average figure is not a statement about the whole league. It is a statement about this pitch against a group of pitches thrown at a similar speed from a similar release point with a similar extension. A pitcher who throws harder is being compared with other pitchers who throw that hard, and a pitcher with unusual extension is being compared with others who share it. The comparison is fairer for it, and it is also narrower than the phrase "above average movement" suggests.

The glossary is also explicit that the sign of a good result flips with pitch type. For breaking balls, offspeed pitches and sinkers, more vertical movement against average means more drop. For four-seam fastballs, better vertical movement against average means more rise. The same column means opposite things in adjacent rows of the same table.

1. 1The hand leaves the ballThe tracking system records velocity, release point, extension and the total rotation of the ball at that instant. This is the only moment spin rate describes.
2. 2Total spin is reportedA count of revolutions per minute, with no information yet about direction or usefulness. Published as the release spin field.
3. 3The axis is recordedAn angle in the two-dimensional plane, where 180 degrees is a pure backspin fastball and 0 degrees a pure topspin curveball. This fixes the direction any deflection will take.
4. 4The ball flies and is trackedAir pressure differences push it perpendicular to both its travel and its spin axis, gravity pulls it down, drag slows it, and the seams do whatever the seams do.
5. 5Observed movement is measuredHorizontal and vertical displacement, in feet in the raw data and reported in inches, with gravity left in so that the figure matches what a person watching would say they saw.
6. 6Active spin is inferredComparing the movement that occurred with the movement the axis predicts gives the share of the rotation that was doing work. The remainder is gyroscopic and moved nothing.
7. 7A comparison group is assembledOther pitches of the same type thrown within two miles per hour, half a foot of extension and half a foot of release point. This defines what average means for this pitch.
8. 8The number is publishedMovement above or below that group, in inches, where more vertical movement means drop on a curveball and rise on a four-seam fastball.

The chain that turns a release into the numbers on a public leaderboard. Only the first two steps are measured directly; everything after that is computed.

The public data has a join in it

Anybody who pulls spin figures across a long span of seasons is quietly stitching together the output of more than one measuring instrument, and the documentation says so.

Savant's field notes carry the tell. The velocity column is annotated with the remark that values from 2008 to 2016 came via the earlier tracking system and were "adjusted to roughly out-of-hand release point", which is an honest admission that the older numbers were converted rather than measured on the same footing. A spin direction column sits in the same documentation marked as a "deprecated field from the old tracking system", left in place so that old queries still run rather than because anybody should use it.

None of that makes the historical data worthless. It makes it a different measurement of the same quantity, in the sense that two thermometers of different design will both tell you the room is warm and will not agree to a tenth of a degree.

The practical rule is unglamorous. Compare spin rates within a tracking era freely, compare across one with caution, and never let a small difference across the join carry an argument. That is a general property of tracking data rather than a quirk of spin, and it applies with equal force to every number that changed when the cameras did.

Why clubs trust spin before they trust results

There is a reason a pitching department will make a decision on a spin profile after a handful of appearances while refusing to make the same decision on an earned run average after two months, and it is a sample size argument rather than a fashion.

Spin rate is measured on every single pitch. A reliever who throws a few hundred pitches has a few hundred observations of his own release, each one a direct reading of the same physical act. The figure settles quickly because it is describing something the pitcher does the same way every time, and departures from it are informative precisely because it is normally so stable.

Outcome statistics accumulate at a vastly slower rate. Balls in play arrive a handful per appearance, they are converted into outs by fielders the pitcher does not control, and the results are filtered through opponents, parks and luck before they reach the box score. That is the whole reason the fielding-independent family of pitching statistics exists, and spin data goes a step further in the same direction: it skips the outcome entirely and measures the input.

The corollary is the one that gets clubs into trouble. A stable measurement is not automatically a valuable one. A pitcher can have a beautifully repeatable spin profile and still be hittable, because rotation is only one of the things that make a pitch difficult, and the ones it leaves out are the subject of the next section.

What spin rate cannot see

Command. The number describes what happened to the ball, not where it went. A pitch with ideal rotation thrown to the middle of the plate is a mistake with excellent physics. Nothing in the spin data records the intention, so nothing in it can record the miss.

Deception. Two pitchers with identical release characteristics can be picked up very differently by a hitter, depending on what the arm does before the ball appears and how early the hand becomes visible. That is a property of the delivery rather than of the ball.

Tunnelling. A pitch is hard to hit partly because it shares a path with another pitch for long enough that the hitter must commit before the two separate. That relationship lives between pitches and cannot be read off either one of them individually. Two pitchers with the same fastball spin profile can get completely different results from it depending on what else they throw and how similar it looks for the first fifteen feet.

Sequence and count. The same pitch in a different count is a different pitch to the hitter, and the tracking data is indifferent to the situation it was thrown in.

What the catcher does with it afterwards. A pitch at the edge of the zone becomes a strike or a ball partly because of the hands that receive it, which is a separate skill with its own body of measurement and no relationship whatever to the ball's rotation.

Spin rate is a clean measurement of one physical property. It is not a model of pitching, and the temptation to treat it as one is strongest exactly where the data is cleanest.

Velocity, extension and the reason raw rpm is not comparable

Spin does not arrive independently of speed. The same delivery thrown harder tends to produce more revolutions, because the same finger action is applied to a ball leaving faster. Comparing the raw rpm of a pitcher who throws hard with one who does not is therefore comparing two things that were never on the same scale.

The standard correction is to divide spin by velocity, producing a ratio that asks how much spin this pitcher generated for the speed he threw at. It is a crude normalisation and it does the job, in the same way that dividing anything by the thing it scales with does the job. A pitcher with a high ratio is getting more rotation out of each mile per hour than his peers, which usually indicates finger pressure and release quality rather than arm strength.

Extension belongs in the same conversation. Savant defines effective speed as a "derived speed based on the extension of the pitcher's release", which is a way of saying that a pitcher who releases the ball closer to the plate gives the hitter less time even at the same measured velocity. Extension shortens the flight, and a shorter flight means less time for any deflection to accumulate. A high-spin fastball from a pitcher with enormous extension is doing two things to the hitter at once: arriving early and finishing higher than expected.

The interaction runs the other way too. Because the movement comparison group is defined partly by extension, a pitcher with unusual extension is compared with the small population who share it, which can make his movement figure look less remarkable than the effect it has on hitters. This is one of several reasons the public numbers and the coaching-room numbers can disagree without either being wrong.

What each pitch actually wants from its spin

The single most useful reframing of this subject is that spin is not a virtue, it is a resource, and each pitch type spends it differently.

The four-seam fastball wants a lot of spin, high efficiency and an axis near pure backspin. Every one of those pushes in the same direction: less drop than a hitter's eye predicts from that velocity. This is the pitch that made spin rate a broadcast statistic, and it is the reason the top of the strike zone came back into fashion. A fastball that resists gravity is worth throwing where gravity is expected to have taken it.

The sinker or two-seamer wants the opposite tilt. The axis moves away from pure backspin towards the side, converting some of the vertical push into horizontal run and reducing the resistance to gravity. Spin rate matters less here than where the axis points.

The curveball wants spin on a topspin axis, so that the Magnus force and gravity work together rather than against each other. MLB's glossary states the effect directly: high-spin curveballs generate more break. This is the one pitch where the naive reading of the number is roughly correct.

The slider is the awkward one, because sliders live across a wide range of axes and a wide range of efficiencies. A gyro-heavy slider is deliberately inefficient. The rotation is largely wasted by design, the pitch moves relatively little, and its value comes from arriving on a fastball's path and then not doing what a fastball does. Judging that pitch by its rpm figure is close to meaningless.

The changeup and splitter want less. MLB's own definition says lower spin rates on those pitches create increased downward tumbling. The point of the pitch is to look like a fastball out of the hand and then fall out of the sky, and spin that resists gravity is working against the purpose.

Set those five side by side and the reason a league-wide spin leaderboard is close to useless becomes obvious. It is ranking five different jobs on one scale that only one of them is trying to maximise.

The foreign substance rules, and why grip is a spin story

Friction between the fingers and the ball is what converts arm speed into rotation. A pitcher whose fingers grip better imparts more spin at the same effort, which is the entire reason grip enhancement has been a live issue in the sport rather than a footnote about tidiness.

The rules are not vague about it. The Official Baseball Rules prohibit a pitcher from applying a foreign substance of any kind to the ball, from defacing the ball in any manner, and, in Rule 6.02(c)(7), from having "on his person, or in his possession, any foreign substance." The comment attached to that clause extends it to attachments: a pitcher may not attach anything to either hand, any finger or either wrist, and the umpire decides whether such an attachment is a foreign substance for the purpose of the rule.

The penalty is set out separately and it is not a warning. For violations of the relevant clauses, the rules provide that the pitcher "shall be ejected immediately from the game and shall be suspended automatically", with the minor league version of that automatic suspension specified at ten games. A separate rule, 3.01, covers any player who intentionally discolours or damages the ball by rubbing it with soil, rosin, paraffin, licorice, sandpaper, emery paper or another foreign substance, and attaches removal from the game plus an automatic ten-game suspension.

One substance sits inside the rules rather than outside them. Rule 4.01 requires the umpire to ensure "that an official rosin bag is placed on the ground behind the pitcher's plate prior to the start of each game", and the comment to Rule 6.02(d) states that a pitcher "may use the rosin bag for the purpose of applying rosin to his bare hand or hands." The same comment forbids dusting the ball with it, applying rosin from the bag to a glove, or dusting any part of a uniform with it. The permission is for the hand and nowhere else.

There is also a discretion clause worth knowing about, because it explains why enforcement looks inconsistent from the stands. Where a pitcher's actions do not change the characteristics of a pitched ball, the rules allow an umpire to warn rather than apply the penalty for the relevant clauses, and to apply the penalty if the pitcher persists.

The thing to take from all of this is the causal chain rather than the enforcement history. Grip determines friction, friction determines how much of the arm's energy becomes rotation, and rotation determines movement. A rule about what a pitcher may carry in his glove is, in physical terms, a rule about spin rate, and a change in how strictly it is enforced is a change to the conditions under which every spin figure in the sport is produced. Anyone comparing spin numbers across a period when enforcement changed is comparing measurements taken under different rules, in the same way that comparing offensive statistics across a change to the strike zone compares two different games.

Five things to check before you quote a spin rate

The pitch type. A number with no pitch attached to it cannot be interpreted. The same figure is excellent on a curveball and irrelevant on a slider.

The efficiency. Total spin without active spin is a count of revolutions of unknown usefulness. If the efficiency figure is not available, the rpm figure is a partial reading and should be quoted as one.

The axis. Direction of movement is set here and nowhere else. Two pitches with identical rate and identical efficiency and different axes are different pitches.

The velocity it came with. Spin scales with speed, so a raw comparison across pitchers who throw at different speeds is not a like-for-like comparison. Dividing by velocity fixes most of it.

What the pitcher is trying to do. A low spin rate on a splitter is not a weakness, and a high one on a changeup is not a strength. The question is never how much spin, it is whether the spin is pointing where the pitch needs it.

Get those five in front of you and the rpm figure stops being a score and becomes what it always was: the first of several inputs, useful mainly for what it lets you compute next. That is the same discipline the rest of the tracking data on the sport demands, whether the number in question describes a pitch leaving a hand or a ball leaving a bat, and it is why the pitchers who benefit most from this data are the ones being coached with all four numbers rather than the one that fits on a graphic.