Thanks for sharing =]
12 runs isn't a huge sample, but at the same time, the probability of pure RNG following your theory is extremely low (isn't it 1 chance on 4096?).
Interesting, i'll post the sequence of my next run, assuming it doesn't take me ages to get the statue i'm missing.
Correct
It's 1/2
for the chance that order A would produce one jewel and order B would produce the other jewel
It's 1/2
for the chance that order A would produce light jewel and order B would produce magic jewel
So it's 1/2048 for the double tail end assuming the null hypothesis that the order of the first two colossal ancient kills is irrelevant to the jewel produced