Is number of followers a really good metric?

[CaioHVectorA]

Based on thread: Bsky thread

And some absurd values:

i dont stopped yet to analysis the code deeply, is it based on a calc with followers_count / following_count ?

[CaioHVectorA]

Testing with some big accounts:

5M Followers | 2 Following

300k Followers | 500 Following

[CaioHVectorA]

Maybe a normalization or a limit will be a good one.

[rdenadai]

I open a PR to avoid this explosion, and a tentative to normalize the final result as a percentage.
The value calculated by the initials rules has a greater weight and the followers/following dynamic has a small weight (it can duplicate the original value).
That does affect the final score, but has small amount of weight in overall ...