Shield Attestation Question (probably a dumb one)

[Garret]

This could very well belong in the newb forums, but I am going to ask it anyway. What I am actually looking for is for someone to tell me I am a superstitious moron, but I'd at least like clarification on that part.

Ok, Shield Attestation is a bit strange in comparison to the others because it always drops with one other Attestation. That being said, the one time we've gotten it before was in helping a JP Pld on our server get Aegis, when doing it he asked that we kill ALL the Hydras with Goublefappe first in a specific order, I believe the order was Rdm > War > Pld & he had this was because by killing the Pld Hydra last it makes Shield Attestation drop.

Now, the superstition side is that we were 0/3 before doing that strategy, and then the one time we did it it worked, and we're 0/2 since that time when not using it. To try & 'test' his theory I did the same thing on another Attestation NM, killing all the Hydras before the NM in a specific order & seeing if the hydra mob set I killed last would result in that attestation dropping. It didn't work, however, it could be 'argued' that because Shield is different, since it always drops with another attestation, that something could apply to getting it to drop which does not apply to other attestations.

So, could someone tell me I'm just flat out retarded please so I can not worry about so ridiculous a 'factor' on our farming run tonight.

[Seyrr]

Sure, no problem. You're retarded. :D

Originally, Goublefaupe dropped three attestations: sword, great axe, great sword, and had an equal chance of dropping each one. Same as all of his other Attestation buddies. When they added Shield, they needed to add it to one of the 5 NMs, and the PLD one made the most sense. But they didn't want to make it less likely that you'd get one of the other attestations, since it wouldn't make sense to have a 1/4 chance for those and a 1/3 for all other attestations.

So Goublefaupe's drops work as such:

One of (sword, great sword, great axe) with a 1/3 chance for each.
1/3 chance of shield, completely separate from the others.

You've had six runs and one dropped? A simple binomial distribution for 6 runs gives us an expectation of 2 drops, with a standard deviation of about 1.155. So 1 drop is within a single standard deviation - it's really not that weird at all. But people really want to find a pattern so they can feel like they have some control over it... because really, it is frustrating to make run after run after run with no success.

[Garret]

But people really want to find a pattern so they can feel like they have some control over it... because really, it is frustrating to make run after run after run with no success.

Yeah, I am fully aware of the wanting there to be a pattern in making it drop thing, my chief frustration was that the one time we do something wicked crazy, we get the drop. I blame gremlins.

[Seyrr]

But people really want to find a pattern so they can feel like they have some control over it... because really, it is frustrating to make run after run after run with no success.

Yeah, I am fully aware of the wanting there to be a pattern in making it drop thing, my chief frustration was that the one time we do something wicked crazy, we get the drop. I blame gremlins.

Oh, you'll find no argument here.

As a side note, a good model for how many runs it should take to drop... let's use a geometric distribution with p = 1/3. The geometric distribution asks "How many trials does it take before we have a success?" Then we get:

Expected value = 3
Variance = 6
Standard deviation = about 2.45

So a quick and dirty estimate gives us that about 68% of the time, you'll get it within 5 runs, and a 95% chance you'll get it within 8.

I could do a hypergeometric distribution to ask the question of how long you expect it to take to get two shields, and so on... but I think my boss is going to notice if I do much more. So that's all you get from me right now.