The second problem was its shape. The object was too far away to image directly, so its shape had to be inferred from how its brightness changed as it tumbled. And the brightness varied enormously, by a factor of about ten, on a cycle of roughly eight hours.


