In a lot of situations, the observation is that the current favorite theory doesn't work for some observation, but the new theory explaining it only comes much later.