Your ability to miss the point completely and say something random is awesome.
That is exactly why you dont want to insert another RNG factor such as /roll in already huge RNG factor such as chance for an item to drop.
If at any point the item dropped, I would be the one getting it because I had most DKP since I dont need anything else.
That system would let me bis my character the fastest way because I could lose the roll against A LOT of people rolling on the same trinket way too many times.
But I would suggest trying to join a guild with working DKP system and stay in it for a long period of time before commenting again.
Do I have to repeat my question for the third time?
Out of 300 rolls, where you can get a number from 1-100, why did I NOT get the numbers: {34, 44, 64, 89} at all, and I got number: {3} 6 times??
BECAUSE SAMLE SIZE IS SMALL.
Out of 15 boss kills where you have 15 different items, one can assume that every 15th kill would result in a drop of 1 out of 15 items that you need.
What if the item drops 2 times in a row after 28 kills?
What if the item drops 5 times in a row after 70 kills?
It will still get in the 1 out of 15 chance to drop by those numbers.
This is EXACTLY why you can NOT link statistic and probability with things in wow such as:
CHANCE FOR AN ITEM TO DROP + CHANCE FOR YOUR ROLL TO BE HIGHER THAN SOMEONE ELSES ROLL.
There is a reason people dont see items for a long long loooong time, even tho they already killed the boss more than double the times what statistical "laws" would call a "100% chance to drop next time you kill it", by your logic.
There is a reason why I farmed Aldrianas every week for 6 months on my warrior, coined it after so much, just for it to drop the next raid I joined on my warrior.
I can keep on giving examples like this where the statistics and probability you value so much in these conversations fail in the game.

