A bead sits on a wire at the top left. Let it go, and gravity carries it to the bottom right. Which shape of wire gets it there quickest? Not the straight line: the shortest path is not the fastest. Not the circle Galileo guessed, though he was close. The winner is a curve that dives steeply, dips below the finish, and comes back up to meet it, and it wins against every slide you can draw.
For the curious kid.
A bead only goes fast once it has fallen. So a slide that drops steeply at the start builds speed early, and then spends that speed racing along the rest of the way. Drop too steeply, though, and the slide gets long: you have to come back up to reach the finish. The perfect slide balances the two, and it turns out to be a curve with a name: the cycloid, the path a spot of paint on a bicycle tyre traces as the wheel rolls. Draw any slide you like on the wire above and race it. Make it steep, make it gentle, make it wiggle. Every slide you draw will lose to the gold curve, not by a little on a bad day, but always, by a law.
In June 1696, Johann Bernoulli printed a challenge to "the sharpest mathematicians in the whole world": find the curve of quickest descent. He gave them six months, then extended it to a year. Isaac Newton, then running the Royal Mint, received the problem one afternoon, solved it that evening, and sent his answer in without a name on it. Bernoulli knew anyway. Tanquam ex ungue leonem, he wrote: "I know the lion by his claw."
Deeper.
Bernoulli's own solution is one of the loveliest arguments in physics. Light, said Fermat, takes the path of least time, and it bends at a boundary according to Snell's law. So imagine light travelling through a stack of layers in which its speed grows like the falling bead's, v = √(2gy). Snell's law in every layer then says sin θ / v is constant, and that single condition, written out, is the equation of a cycloid. The bead behaves like a ray of light in a world whose optics are gravity. The problem's name, brachistochrone (brachistos, shortest; chronos, time), was Bernoulli's; the general method it provoked, finding the best curve among all possible curves, grew into the calculus of variations, and from there into Lagrange's mechanics and, eventually, the principle of least action that underlies all of modern physics.
The cycloid keeps a second secret, found by Christiaan Huygens twenty years earlier: it is a tautochrone. Release beads from anywhere along its descending arc and they reach the bottom at the same instant: a bead starting near the bottom dawdles, a bead starting high plunges, and the times agree exactly (π√(R/g), independent of where they began). Press "Let go together" and watch six beads from six heights arrive as one. Huygens used this to design a pendulum clock whose beat would not depend on how widely it swung.
Honesty notes, museum policy. The bead here is frictionless and starts from rest; its speed at any point depends only on how far it has fallen (energy is conserved), so along each straight piece of a wire the travel time has an exact closed form and the page adds those up: no numerical integration, no time step, no drift. The gold curve is a true cycloid through the two points, its time checked headless against Bernoulli's θ√(R/g) to nine digits, and the drawn slides are resampled to 240 straight pieces. A slide that rises above the start is reported honestly: the bead stops. Galileo, in 1638, proved that a circular arc beats the straight line and guessed it was the best possible; here the circle loses to the cycloid by only three per cent. A good guess, and wrong.
The quickest way down is not the shortest. It is the one that learns to fall first.