I posted about this topic a while back, but now i like to get an answer from the devs themselves. My theory is that you do not use a true 6 sided dice rng for determining the result of pyramid type rolls (and probably other special missions as well).
Background: as i reported previously in this forum, i seem to have amazingly bad luck with getting a noxmid. Right now, i have tried 30 times to get a noxmid (took me more than 15 keys and 3 weeks). Statistically, the chance of this much bad luck is (5/6)^30=0,29%, assuming there is no correlation between different draws. But one thing i noticed: chances of getting the same pyramid type twice within one key unlock period seem to be very high. So it seems there is correlation. Which is very strange and unfair in my opinion. So what else could be strange about your rng?
Well, I suspect that the chances are also altered (lowered) based on demand: in my case, i have two monsters waiting for super evolution, both need noxmid. Which happens to be the only type i am Not getting. If this is true, it would also seem very unfair to me.
Can you confirm all this? Obviously other factors are involved except true rng. Which factors, and why did you make this decision?
Ive been told by the devs that it is completely random.
When something is completely random, there will be extreme cases on both sides of it. You just have the bad luck to be on the bad side of rng. someone else have gotten exactly the ones they need every time.
I know about the nature of randomness, that is exactly why i find it hard to believe that there no other factors involved.
Anyway, thank you for your answer. I believe that the devs told you that it is completely random, however i choose to not believe their statement based on my own experience, and more importantly, my knowledge of statistics.
Computers can’t actually generate anything random - they can only fake randomness using algorithms (creating pseudo-randomness).
In the realm of computers, not all randomness algorithms are created equal. It is possible that they are using a relatively low quality randomness algorithm.
I believe they use Unity, and I know that Unity’s Random class is more optimized for performance and less for quality randomness. Better results may be obtained from better algorithms.