Do you have any more rigorous documentation than this? Because 0 for 30 is within acceptable parameters as being brushed off as “bad luck” if the given rate was 10%. But 0 for *120* would be about a 1 in 300,000 cumulative probability. 2 (or fewer) for 300 should be a 1 in 90 *billion* (yes, billion with a b with a b) cumulative probability for a stated rate of 10%. Without hard data, this is going to be written off as “anecdotal”, though.

Something as rare as this (EVK only has a *stated* drop rate of 10% of a normal vault key, which is in and of itself only about a 10% rate on vault weekends) is going to be *really* hard to get a trend of. But you can show trends with *relatively* small sample size, especially if your samples account for only drops where a Vault Key/EVK was dropped.

For example, lets say hypothetically 100 rigorous samples were taken where some kind of vault key were dropped. The stated rate of epic vault keys is 10%, but only 3 epic vault keys dropped out of the 100. The cumulative probability of getting 3 (or fewer) EVK’s out of 100 is about 0.78%. If this were on the dev side sampling rates before release, this would *definitely* be in the realm of need more samples. On the outside looking in, this can still be described as “bad luck” because we have a small sample size. But to accept this we have to also accept that it is *unlikely* for a mistake to have been made and released in that state. If we take the tiniest leap in logic and assume the intended rate is 1%, and run the numbers again, the cumulative probability of getting 3 (or more) EVK’s out of 100 is about 7.9%. This is not *proof* of anything one way or the other, though.

But what about 0 for 100? Well, if I were getting this sample at *any* point during the dev side, I’d probably be ready to examine the mechanism then and there. If we get this on *our* side, we’ve hit what should be a 1 in 37,000 occurrence already. This is getting less and less likely to be able to be written off as “bad luck”. If we make the same logical leap as we did before and assume a simple typo was left uncaught with inadequate testing (intended rate of 1%), we’d be looking at a 36.6% occurrence. Still, we’d have to make an *assumption* here.

We can look at other samples sets posted in this thread, for example, we have a 1 for 13. That would seem to be almost right on the money for a 1 in 10 rate, at about a 74.5% cumulative probability to have gotten 1 or more epic vault keys by this point. But even at a 1% chance, you would have had a 13.2% cumulative probability to have earned 1 or more epic vault keys. So we’ve had way less extreme outliers if we assumed 1% instead of 10%, but again, from an *objective* standpoint and based on *rigorous* data we can’t really do that…

So, as of now, the *statistical* data we have is inconclusive. It doesn’t help that this thread began as an off the cuff complaint backed only by a 0 for 15, which is only stated way deeper in the thread (and even that, given that the stated rates are correct, will still be experienced by one in five players who get this many vault keys), and making claims that can’t be backed up like “nobody has found them” (then of course easily and instantly rebutted by someone who did) which is likely to just get ignored. This would be infinitely easier for a rigorous check to be run on their end to verify the rates (along with internal check to verify that the 10% number is what should have been posted in the first place) on their end, but such a thing is unlikely to happen until after the weekend, when it is already “too late”, at least for this event. If you (anyone reading this) care at all, you should probably start tracking your drops so any potential mistakes can be identified or we can get enough data to show the rate does indeed trend toward 10% over large samples. So far I’m 0 for 0, so no help there.

Personally, I’m less concerned over “compensation” and more concerned over things being correct, especially with specific drop data continued to be withheld.