Sure, the argument that since ML revolutions worked in the past, they will work in the future is overly simplistic. But I don't think it is irrelevant, because there were (very likely) reasons they worked. I would claim the reasons that, for example, democratic centralism and the vanguard party worked are still relevant in the imperial core, and you would need a pretty strong argument to reasonably suggest abandoning them based on their past success. At the very least, a modern revolutionary strategy should have an answer to the problems previously answered by these strategies.

No strategy will ever have been tested in the exact conditions it will be used in, because all conditions are unique in some way. It will always be necessary to extrapolate from past successes and failures.