In both cases I'm talking about 2 times merged servers, so both are established servers. One merged last year (25k accounts), other merged like 3 years ago (10k accounts).

Surprisingly the latter works a lot more with fillers than the first. So the functioning server has a lot more "played" accounts (doesn't matter main or feeder) than the non-functioning server, not only stats-wise but also being actually used and not only as filler.