Why Does a Curveball Curve?

This is part eight of Lunch Break Physics, the series that has argued a suction cup isn’t sucking, a fridge doesn’t make cold, the sky isn’t blue because it reflects the sea, ice floats because water freezes into a roomy cage, planes fly by throwing air downward, that the shower curtain attacks you for reasons nobody can fully pin down, and that noise-cancelling headphones make silence by adding sound. This week the ball bends, and it bends for an honest reason. Standard physics below; sources at the bottom; corrections very welcome in the comments.

The story you were told

Ask why a curveball curves and you’ll get one of two answers, and they’re both a dodge. The first is that the spin makes the air “grab” the ball and pull it around. That sounds like an explanation but it isn’t one, because it never says which way the ball goes or why, and “grab” isn’t a force air knows how to apply. The second answer is more confident and, oddly, more popular: that the curve isn’t real at all, that it’s an optical illusion, a trick your eyes play as a fast ball crosses your field of view.

There’s a grain of truth buried in that second one, and we’ll dig it out later, because the sharp break really is partly in your eyes. But the ball genuinely moves sideways. High-speed cameras and radar tracking measure the deflection in centimetres; it doesn’t care whether anyone is watching. So neither the vague “grab” nor the flat “it’s an illusion” survives contact with the evidence. The real answer is more satisfying than either, and you can watch it happen.

Watch it happen

Below is a top-down wind-tunnel view: air streams past a spinning ball from the left. The blue side is where the dragged-around air runs with the stream, so the flow is faster and the pressure lower; the red side is where it runs against the stream, slower and higher pressure. The green arrow is the resulting push, and the green flight path is where the ball actually goes. Now press Flip the spin. Watch the blue and red sides swap, the push reverse, and the flight path bend the other way. Then turn the spin rate and speed up and down and watch the curve grow and shrink.

The flip is the whole argument. If the curve were the air randomly grabbing the ball, or a trick of the eye, reversing the spin wouldn’t cleanly reverse the curve every single time. It does, because the direction of the curve is locked to the direction of the spin through the pressure difference the spin creates. Nothing about the ball changed except which way it turns.

The actual physics

Start at the ball’s surface. Air is slightly sticky (it has viscosity), so the very thinnest layer of it clings to the ball and moves with the surface. This clinging layer is the boundary layer, and on a spinning ball it gets carried around in the direction of the spin, dragging some of the neighbouring air with it [1]. That’s the honest version of “the air grabs the ball”, except it’s the other way round: the ball grabs a little air.

Now add the ball’s motion through the air, or equivalently, a stream of air flowing past it. On one side of the ball, the air being dragged around by the spin points the same way as this oncoming stream, so the two add up and the flow there is faster than the free stream. On the other side, the dragged air points against the stream, the two partly cancel, and the flow there is slower [1]. You now have fast-moving air on one side of the ball and slow-moving air on the other.

Fast-moving air has lower pressure; slow-moving air has higher pressure. (That’s the same Bernoulli trade-off that pulls a shower curtain inward and helps hold a wing up.) So the spin has manufactured a pressure difference straight across the ball: low on the fast side, high on the slow side. Pressure always pushes from high toward low, so the ball is shoved sideways, toward the fast, low-pressure side, and it keeps being shoved the entire flight [1]. That sideways shove is the Magnus effect, and its direction is set by the spin: reverse the spin and you swap which side is fast, which swaps the pressure difference, which flips the curve. Exactly what the demo does.

That fast-side, slow-side account is the honest half of the story, and there’s a deeper half worth one paragraph. Because the air is sticky, the boundary layer doesn’t just change speed on the two sides, it lets go of the ball at different points. On the side sweeping along with the airflow the layer keeps its momentum and clings on longer, so it separates late; on the side sweeping against the flow it is slowed and peels away early. With the layer hanging on longer on one side, the whole wake behind the ball gets flung off to that side, and by Newton’s third law the ball recoils the other way — toward the faster, lower-pressure side the Bernoulli picture already pointed at [7]. Pressure and deflected-wake aren’t rival explanations; they’re the same event, once counted with a pressure gauge and once with a momentum budget, exactly as with a wing. This asymmetric-separation view is also why a smooth ball can misbehave: change how the boundary layer trips on each side and you can even get an “inverse” curve, which is a story for another day [7].

There’s a neat bookend hiding here. Back in episode one a suction cup held on because the air outside pushed harder than the air trapped underneath, a pressure difference across a boundary. In episode five a wing rose because it turned the air downward and rode the pressure difference between its faster top and slower bottom. A curveball is that same story rotated ninety degrees: a moving object, a faster flow on one face than the other, and a pressure difference that pushes it. Pressure and flow, again.

A little bit of math

NASA’s ideal model treats the spinning ball as a stack of spinning cylinders and adds up the lift on each with the Kutta-Joukowski theorem. You don’t need the calculus; you need what falls out of it. The sideways force on the ball works out to

$$L = \tfrac{4}{3},\pi^{2}, b^{3}, s, \rho, V$$

where b is the ball’s radius, s the spin rate, ρ the air density, and V the ball’s speed [2]. The shape of the equation is the point. The force grows with the spin rate s and with the speed V: spin it faster or throw it harder and it bends more, which is precisely the two sliders in the demo. It also scales with air density ρ, and that isn’t just decoration, it’s why a curveball breaks less in thin mountain air. This is an idealised figure (it assumes smooth, non-viscous flow and then quietly relies on viscosity to create the spin in the first place), so real pitches carry a measured correction factor [2]. But the proportions are honest: more spin, more speed, denser air, more curve.

The same trick, elsewhere

Once you can see the Magnus effect, it turns up everywhere a spinning thing moves through air. Each of these runs on the exact mechanism above.

• The soccer “banana” free kick. A struck ball given sidespin bends around the wall toward its low-pressure side; the famous impossibly-curling free kicks are pure Magnus, just with a bigger, slower ball [3].
• The golf drive that hangs in the air. A clean iron shot puts backspin on the ball. Backspin makes the top surface run with the airflow, so the low-pressure side faces up, and the Magnus force points upward, adding lift that stretches the carry [3].
• Table tennis and tennis topspin. Topspin points the Magnus force downward, so the ball dives onto the table or dips inside the baseline faster than gravity alone would bring it down. Same effect, aimed at the ground.
• Cargo ships with spinning metal chimneys. A “Flettner rotor” is a tall cylinder spun by a motor; as wind blows across it, the Magnus effect generates thrust at right angles, exactly the Kutta-Joukowski lift from the math above. Anton Flettner sailed a rotor ship across the Atlantic in the 1920s, and modern cargo ships have bolted them on to cut fuel [4].

False friends

These all look like a curveball’s cousin but run on different physics. This is where a sharp-eyed fan should look first.

• The knuckleball. It flutters and darts far more wildly than a curveball, so it seems like Magnus turned up to eleven. It’s the opposite: it’s thrown with almost no spin, and its erratic dance comes from its raised seams tripping the airflow unevenly, not from a steady pressure difference. No spin, no Magnus [5].
• The “sharp break” itself. Here’s the grain of truth in the illusion story. The ball’s real path is a smooth, gradual arc, but batters swear it “breaks” suddenly near the plate. That sudden break is a genuine perceptual illusion: as the ball crosses from your central vision to your periphery your brain misjudges its speed and location, so a smooth curve is misread as a late jump [6]. The curve is real; the snap is your eyes.
• A pitch that just drops. Some of a ball’s downward fall is simply gravity acting over the flight, the same parabola a thrown rock follows. That vertical drop looks like “movement” but needs no spin at all; only the extra drop beyond gravity is Magnus.
• A ball shoved by a crosswind. A gust can push a fly ball off line, and it looks like a curve, but that’s the ball riding a moving mass of air, not a pressure difference its own spin created. Kill the wind and the deflection goes with it.

Fun consequences

So the next time a curveball buckles someone’s knees, you can enjoy the plain mechanics of it: spin, a dragged skin of air, a faster flow on one side than the other, and a pressure difference leaning on the ball all the way to the plate. It isn’t the air grabbing it, and it isn’t only your eyes. It’s the same thing that holds a suction cup and lifts a wing, pointed sideways: air pushing harder on one side than the other.

Lunch Break Physics runs Tuesdays at noon. Last week: how noise-cancelling headphones work. Got an everyday-physics puzzle you’d like poked at? The comments are open.

Sources

• [1] Ideal Flow Around a Spinning Ball / Ideal Lift of a Spinning Ball — NASA Glenn Research Center — a thin boundary layer of air sticks to the ball’s surface and is entrained in the direction of spin; adding this to the free-stream flow makes the net velocity greater on one side of the ball and less on the other, altering the pressure field so a force is generated perpendicular to both the flow and the spin axis, directed toward the low-pressure (faster) side.
• [2] Ideal Lift of a Spinning Ball — NASA Glenn Research Center — applying the Kutta-Joukowski theorem (lift per unit length L = ρ·γ·V) to a stack of cylinders and integrating gives the ideal lift on a spinning ball, L = (4/3)·π²·b³·s·ρ·V; the model is an idealised, non-viscous flow, so real balls carry an experimentally determined lift coefficient because viscosity, the boundary layer, and stitches make the true flow complex.
• [3] Aerodynamics of Baseball / sports aerodynamics — NASA Glenn Research Center — aerodynamics governs the flight of balls across many sports including baseball, golf, and football; the same spinning-ball lift that curves a pitch produces the curve of a struck soccer ball and the added carry of a backspun golf ball.
• [4] Flettner rotor — Wikipedia and Flettner rotors — IMO GreenVoyage2050 — a Flettner rotor is a spun vertical cylinder that generates an aerodynamic lift force via the Magnus effect (Kutta-Joukowski lift) at right angles to both the wind and the rotation axis; rotor ships from Anton Flettner’s 1920s Buckau to modern cargo vessels use them as motor-driven sails to cut fuel.
• [5] Aerodynamics of Baseball — NASA Glenn Research Center — the raised stitches on a ball stick up out of the boundary layer and disturb both it and the free-stream flow; on a near-spinless knuckleball this seam-driven, asymmetric, unsteady flow, rather than a steady Magnus pressure difference, produces the erratic motion.
• [6] Transitions Between Central and Peripheral Vision Create Spatial/Temporal Distortions — Shapiro et al., PLOS ONE (2010) and The Vertical Illusions of Batters — SABR — a curveball’s trajectory is a smooth, continuous arc, but the perceived sudden “break” (and the “rising” fastball) is a perceptual illusion arising when the ball moves from the batter’s central to peripheral vision, causing a misestimate of its speed and position.
• [7] Magnus effect — Wikipedia — the spin makes the boundary layer separate asymmetrically, so the wake is deflected to one side (in the direction of spin) and, by Newton’s third law, the ball experiences a reaction force in the opposite direction — the same force the Bernoulli pressure-difference picture describes; certain smooth-sphere conditions with turbulent flow on one side and laminar on the other produce the reversed “inverse Magnus effect.”

About the Author

Stefan Christoph is a Principal Solutions Architect at AWS, focused on agentic AI, media & entertainment, and helping builders move from demo to production. He writes about AI architecture, developer productivity, and the future of software.

This is a personal blog. Opinions expressed here are my own and do not represent the views or positions of my employer.

Learn more →

• Physics
• LunchBreakPhysics
• Science
• Curiosity