06-09-2016, 01:14 AM
All my statistical info is 2 years old. Anecdotally, I find the 1.9 generated worlds not much different from before in diamond distribution when mining at y==11. Yes, I too have a vague feeling that I am finding more. I am lazy so I have not redone my MCEdit studies. I suspect the statistics that I found by a way more obsessive player have been updated -- you can google for minecraft diamond distribution probabilities. Unless that player found a new obsession 
.
My tendency in this situation is find the decompiled 1.7 source and 1.9 source and look at the diamond generation code for the probability parameters used -- except that is type of thing that I do for work and I prefer to play in MC instead. Besides it is Java code, yuck, or x86 assembly, double-yuck, or JVM bytecode, triple-yuck. And world generation is done in multiple passes which can wipe out a pocket of diamond to make a cave or a lake.
I eyeballed using MCEdit and the obsessive statistician actually measured the diamond distribution in generated worlds both by x-z distribution and y height distribution. He then overlayed mining tunnels and counted the diamonds found against what was there.
.My tendency in this situation is find the decompiled 1.7 source and 1.9 source and look at the diamond generation code for the probability parameters used -- except that is type of thing that I do for work and I prefer to play in MC instead. Besides it is Java code, yuck, or x86 assembly, double-yuck, or JVM bytecode, triple-yuck. And world generation is done in multiple passes which can wipe out a pocket of diamond to make a cave or a lake.
I eyeballed using MCEdit and the obsessive statistician actually measured the diamond distribution in generated worlds both by x-z distribution and y height distribution. He then overlayed mining tunnels and counted the diamonds found against what was there.
/Rowebot