That's not really what a seed does. It's just a starting number of the pseudorandomly generated sequence. In a normal PRNG algorithm this can't produce heavily nonrandom results, as the sequence will be random no matter where it starts. Seeds are just used so that results aren't the same for everyone.
Gungho could be rigging the RNG - I highly doubt it would be the first time companies have done that. We at least know what the rates for collab godfests are, since people collected data for those and afaik there was never a case where drops from the same tier/star rating had a different droprate. I've personally never had anything happen that would be incredibly unbelievable, but my sample size is obviously not large enough to draw any conclusions. You should keep in mind though that randomness can be far more streaky than you'd expect.
As for true RNG, that's largely irrelevant. Good PRNG algorithms have been heavily tested and produce extremely good results.
I agree completely and I'm not explaining myself correctly.
Again - what I'm describing is a randomization based on an algorithm which would be unique for each player.
player A wouldn't be able to roll mob X with any frequency but has no issue rolling Z
player B wouldn't be able to roll mob Y with any frequency but has no issue rolling Z
needed for decent team building X,Y,Z
"large" player samples would be purposely biased - grouped samples of multiple people playing would bias the information because of mixing of the different unique statistical PRNG settings on a single account.
you could manipulate the data to fall within a statistical range (say 1% chance when the data is grouped, but modify it such that in truth you may have a 3% chance of rolling certain cards while another person would roll with a 0.3% frequency)
I think I am using the word seed incorrectly - my feelings lead me to believe that one of my accounts will roll multiples of certain mobs while other accounts will roll multiples of different mobs - with a similar frequency but just not the same cards. where I feel the randomness breaks down with when one of my accounts starts to roll 6-7 of a certain gfe while another account does the same thing, but just with a different mob.
now - is it just coincidence? - maybe, but the game sure feels rigged at times, just like a streaky RNG will - but, given this game is all about math - and the countless number of calculations that occur with each game I have a hard time believing that gungho isn't massaging the data.
I highly doubt that there is a seed per user. Given a seed, the random number generator would always return the same number, so they'd have to store the rng state for each user in order to get new values weeks apart. It wouldn't make sense and it'd serve no purpose. There's probably one or a set on each server (physical machine or VM, not NA vs EU) that are shared across all users. Probably the same ones used to determine which orbs fall and what monsters drop.
- if you could make it difficult to roll for a certain team, then a different REM pattern for each player would be the perfect way to massage the statistics an encourage people to roll more.
- second - gungho is constantly watching what players do, how we spend and what bonuses are needed to keep people playing - look at the reward differences between the servers - it is nothing to think that the couldn't keep individual REM bias on record.
(And I realize how tinfoil hat I'm sounding - so I'll back off - and - say I'm still supporting the game and understand that there is no way I can test this theory and just have to depend on the honesty of Gungho)