Why Witch loot drops suffer so badly - the stats:

"

Cronk wrote:

"

Duskbane wrote:

433 items is nowhere near a large enough sample size to rule out randomness. Anyone who has taken a statistics class can tell you that. Come back at about 433,000 at the minimum.

Yes it is.

No, it's not. Again, ANYONE who has taken even a high-school level statistics class will agree with me here.

http://en.wikipedia.org/wiki/Normal_distribution

EDIT: If that's too complicated for you to get, how about this? Go find a 6-sided die and roll it 433 times. Prove to me that, based on those 433 trials, each side of the die has approximately 1/6 chance of occuring.

It can happen, but...

I doubt you'll get those results.

"

Tabx wrote:

"

Nochalo wrote:

I think i read your previous topic, and remember that I was ashamed of "RNG is RNG" crowd. Surely, it can happen, that first 100 currencies you get will be exalts, but only in theory. The theory is one thing, and actual statistical chances are the other.
That said, I'm not sure, that int gear lacks in drops.
Its not reasonable to divide all ger by 6, 'cause there are 6 classes. You have to look deeper. After that it's obvious, that "pure" classes should have less drops. Why? Because less players can benefit from "pure" gear.
For example - as a mara, i can use everything, sure, but typically, STR gear, or STR/DEX or STR/INT - depends if i follow in the direction of templar or duelist.
But STR/DEX and STR/INT can be used also by duelists and templars. Therefore there should be more general, less "pure" gear in game.

That theory (should) fall apart when the developers realize that many people play by themselves and, in turn, rely on drops. When I read his first post I totally agreed with it. I haven't done any maths like he has, but it's certainly been my anecdotal experience. Hopefully, the developers "look deeper" and see that this game doesn't have an economy worth trading in (that's another post) and many people simply don't. Mixed classes can use pure gear just as much as pures can. I don't see any problem with having an even split for gear.

Actually many pure classes cant use pure gear of another classes due to stat requirements I know on my elemental templar I cannot use alot of dex heavy gear such as rapiers it just wont happen thats why their BIS weapons will always be corsair swords.

"

Bru741 wrote:

"

fsg wrote:

you seem to have misunderstood on fundamental piece of game design here

there are no classes

Wow, try to come off smart but post something utterly retarded like that? Well played son.

When he states classes he obviously means stat-archetypes, like STR-based, INT-based, DEX-based or a hybrid...

is it so obvious? he is talking about "witch gear" and balancing drops towards your "class"

well if i want to build a str based witch using two handed axes, i would be completely fucked in the ass wouldnt i?

"

fsg wrote:

"

Bru741 wrote:

"

fsg wrote:

you seem to have misunderstood on fundamental piece of game design here

there are no classes

Wow, try to come off smart but post something utterly retarded like that? Well played son.

When he states classes he obviously means stat-archetypes, like STR-based, INT-based, DEX-based or a hybrid...

is it so obvious? he is talking about "witch gear" and balancing drops towards your "class"

well if i want to build a str based witch using two handed axes, i would be completely fucked in the ass wouldnt i?

No, you'd be in the pure STR sample group and scoring piles of useful equipment :)

"

Cronk wrote:

Questioning the sample size is just another way of saying "I don't want to discuss this right now", which is fair point, but, in which case, simply don't post if you don't want to discuss the topic for it's real pros and cons. Simply stating "I prefer my drops to be silly-rare" with some justification goes a lot further than a pissing contest about sample size (for which this sample size is perfectly adequate).

No, I will discuss the pros and cons of the system when you have proven your point. You haven't. Go get a larger sample size to back up your information.

I'm no teacher, but if I were, and I gave you a (for simplicity's sake) math test, and you turned it in with only the answers, showing no work, I'd hand your paper back and ask you to try again.

for these numbers to mean anything don't we need to set a baseline of what constitutes "witch gear" being as there are so many possible combinations of gear, passives, gems..etc its hard to really say whats what. Are you considering witch gear something along the lines of "what i need for my spec as a witch" or maybe "anything with intellect spell damage, and ES" If its the former I could possibly sympathize with you especially if your spec is an oddball one that requires harder to find gear but if its the latter all your math is meaningless as I see that kind of gear drop constantly. I mean really your whole post is meaningless because it lacks context.

"

Cronk wrote:

There is no point in creating a larger sample size as the point has already been proven, all a larger sample size would do would be to increase the numbers, not the cause. The cause would still be the same.

Sad to say, but you really do need to take a class on statistics to know that your sample size is far too small for it to be compelling evidence. Come back with a far larger sample, then more people might take this seriously, including myself. You can't build an accurate model off of so few samples, you'll get laughed out of any college level class on the subject.

I take issue with the OP stating that the insufficient sample size argument is one trying to avoid the argument.

Stating insufficient sample size is a very valid argument used in the world of statistics. Based on your sample size vis a vis real world dictates the percentage likelihood that your data accurately reflects what is actually going on in the game.

For example, take what is possibly the most famous stats guru in town, Nate Silver. When he makes a claim based on statistical data, he adds what he thinks is the likelihood that his interpretation of the data is correct based on many factors, including sample size.

That being said, I won't outright claim whether your sample size is large enough or not. What I would humbly suggest the OP do to aid this discussion, is to if possible use a normal distribution calculation (of course with the necessary assumptions) and then give us the output, including the tolerance level (ie: the percentage likelihood it is correct).

Another 3 quick things I'd like to point out that I believe may aid the discussion:

- just because we see evidence weightage in one area that we believe to be true (currency), doesn't mean any similar evidence for another area must be equally conclusive.

- why stop before level 28? I believe the average gamer on these forums would aim for at least level 60 or 70 and that may be more fair. I understand you had 4 screens but drawing a line across the middle of the level seems a little odd given the classes equipment have varying level requirements at different stages

- i'm a lvl 67 templar and i haven't found any sceptre or staff that impressed me so far. i've stuck mainly to daggers (of all things), yet i've found them phenomenal to use. i haven't found the class specific items thing restrictive in the least...

The sample size debate is really funny. You can do some math to determine what sample size is enough. I'm guessing most of you are like me and either don't remember how to do that math or never learned it at all.

Conceptually, a larger sample is always better, but for a certain number of samples, there will be some probability that your results are within some margin of the true values. What the OP failed to address is what those numbers are so we can't really tell if the sample is adequate unless we do the math ourselves.