There was another thread started a few days ago that asked this same question. You had a bad run of luck. I've needed a shoulder piece for two sets. First set I got the shoulder I wanted with the trait and weight I wanted first try. The 2nd set I got the piece I wanted in the weight I wanted 2nd try but had to change the trait myself.
If 100 people spent forty keys you it wouldn't be really unusual to expect three of them to get all of one set.
I used 15 keys to get iceheart on specific chest, got it... 3 times... ON THE SAME WEIGHT... ALL WRONG
Basically 3 days farming as i only have 2 chars, i want Iceheart Medium, and only got it Light, tried to go with random, used 3 keys got Grothdarr, 3 times... ALL ON DIFFERENT WEIGHT... I love and hate my life
furiouslog wrote: »So, for 5 keys, you are basically paying to increase your chances of getting the set you require, and the expected value of the keys relative to a positive hit on your set is marginal at 50/50. I purchased 9 of the Dom/EG coffers, in search of EG shoulders, one after the other, and 8 of the 9 contained Dom shoulders. The probability of that happening if it were a true coin flip is less than 2 percent.
So my question is this: are the loot tables adjusted in favor of sets in some way that is not apparent to me? It’s crazy that I blew 45 keys to get one pair of shoulders that was the right set but wrong type/trait, and I now have a pretty high level of confidence that it’s not random. I had to run 23 vet dungeons on HM to get one piece that I need to transmute and work around anyway.
Anyone else experience something similar when spending multiple groups of 5 keys?
gatekeeper13 wrote: »I had more chance in getting what I wanted from random chests than from specific dungeon chests. Spent 100 keys in SotH and DB chests to get heavy Thurvokun and medium Velidreth shoulders. Got none. Spent less than 30 in random chests. Got both. Never bought a dungeon specific chest after that.
CelticStones wrote: »
furiouslog wrote: »There was another thread started a few days ago that asked this same question. You had a bad run of luck. I've needed a shoulder piece for two sets. First set I got the shoulder I wanted with the trait and weight I wanted first try. The 2nd set I got the piece I wanted in the weight I wanted 2nd try but had to change the trait myself.
If 100 people spent forty keys you it wouldn't be really unusual to expect three of them to get all of one set.
You've reframed the question, but in actuality, it would be odd for that to happen. The probability of getting 8 of the same set on 8 draws is 0.39 percent. I'm not going to build a decision tree, but if 100 people pull 8 sets each, the probability of 3 pulling 8 of the same set is still pretty low. It would be unusual for 1 of the 100 to pull 8 of the same set. It would be improbable for 3 of them to do so.
I do statistics for a living, so I get small sample sizes, but for a sample of 9, that is enough to reject a null hypothesis that the outcome is random at a pretty decent level of confidence. 9 flips gives me a standard deviation of 1.5. The expected value is 4.5 Earthgore shoulders, and two standard deviations puts that at a range of 1.5 to 7.5 Earthgore shoulders. Given that my hypothesis is that the loot tables are adjusted out of favor relative to intention based on available participant information to provide an incentive to go on more dungeon runs, most statisticians with this information would reasonably conclude that the loot algorithm is biased against intention with a level of confidence exceeding 97%. Even if we don't accept my hypothesis, and we are just trying to detect unintentional mayhem, it's better than a 95% confidence level that the result is not random.
I know from personal experience that billions of dollars have been invested by executives who had less statistical confidence to go on, and I think that something smells about this whole thing. When I run vMA over and over with no inferno staff, it's still in the realm of probabilistic outcomes. This Earthgore scenario is improbable enough to be suspicious of the assumption of true randomness. Or maybe I am just super unlucky. But I don't think so. The numbers don't lie.
FrancisCrawford wrote: »But would you have posted this thread if you'd gotten a more balanced result? Do other people who get more balanced results start threads like this? You surely see where I'm going with that line of discussion.