I'm assuming in that thread you linked to, that you were controlling the longbowmen and the AI was controlling the crossbowmen? Either way, from most tests I've done so far, against VH AI, it seems the unit the person is controlling always performs significantly better than the unit the AI is controlling. AI tends to mess about with unneeded movement, reforming, etc. which takes precious time.
I think you'd get the most reliable data if you controlled both units, in turn, and fired on the same enemies, comparing them like that instead of pitting them 1v1 with you controlling one side. Of course AI longbowmen tend to be stupid, often switching to burning arrows :\
Bookmarks