So... gets me wondering. Does the loot server split allocation pathways into parallel routes that can be added to? Or, are all requests dealt with in one processing line that rewrites what is available since the last 'loot injection'?
If the first is true, your entire time in an instance could be with a better or worse RND function and a particular loot not in the pool at all for a time while another player may be on a different pool line that has a loot potentially available.
If the second is true, then all that can be done is put firecrackers up the hamsters' ....s, i.e. up the frequencies.
Anyone tried a deep dive into the secrets sections, and how MA might actually solve the request-flooding issue?