Ptolemy
Can a wrong model still predict correctly?
That time, that place
Alexandria under Rome. In front of Ptolemy lay more than seven centuries of accumulated observations from Babylon and Greece. His task was not to look anew but to bind all of it into a single system of calculation.
Why this question
The planets were the headache. Mostly they move steadily one way, then slow, stop, and for some days go backwards before resuming. If the Earth sits at the center and everything turns in perfect circles, that cannot happen.
What was found
He set a small circle riding on a large one: the planet travels the small circle while the small circle's center travels the large one. Seen from Earth, the planet then appears now and then to go backwards.
To this he added further devices — offsetting the circle's center from the Earth, and measuring uniform speed from yet another point. An elaborate contraption, but the results were striking: it predicted where and when a planet would appear with real accuracy. The premise he started from was wrong, and the predictions were right.
A single perfect circle could account for the heavens
The calculations checked against observations in the Almagest
It stood as the standard for fourteen centuries. Even when Copernicus overturned it, his predictions were at first no more accurate