In 1846 Uranus, the farthest planet known, kept drifting off the path Newton's gravity predicted. Urbain Le Verrier asked what unseen planet would pull it like that, and where it would sit. He worked out a spot in the sky and sent it to Berlin. That night Johann Galle and Heinrich d'Arrest looked, and found a planet within about a degree: Neptune.

Then Mercury misbehaved. Its orbit turned round the Sun a tiny bit too fast, about 38 seconds of arc a century more than Newton allowed, by Le Verrier's count. He tried the same trick: a small planet close to the Sun. In 1859 a country doctor reported a dot crossing the Sun, and the planet got a name, Vulcan. Astronomers hunted for it again and again.

It was never there. Nothing was pulling Mercury; the rule itself was wrong. In 1915 Albert Einstein's new theory of gravity gave Mercury's extra turn, 43 seconds of arc a century, just as measured. No planet needed.

When a forecast misses, something unseen is pulling, or the rule is wrong. The same clue fits both. Only looking tells you which.

The sky is our own drawing; distances and the turn of Mercury's orbit are not to scale.

So, real or not? What did you say?

Sources: Encyclopaedia Britannica, "Urbain-Jean-Joseph Le Verrier"; "That Star Is Not on the Map: The German Side of the Discovery of Neptune", arXiv, 2021; MathPages, "Anomalous Precessions" (with Einstein, 1916); Armagh Observatory and Planetarium, "Vulcan: the solar system's ghost planet".
Sounds: our own, made in code (earshot, CC0).