For the past few months I have kept track of the outcomes of my ring gambles. I have been saving up the calculation until my n was large enough. (Note: nagel and manald were in my stash and I immediately muled away the soj's once they appeared.)

Here are the results of 396 gambles:
328 magical
34 rare
26 set
8 soj

I am testing the hypothesis that the gambling outcome is 85% magical, 7% rare, 5% set and 3% unique. I used the equation for testing a claim about a proportion. At a significance level of 0.05 the alpha is 1.960.

Unique:

((8/396)-0.03)/sqrt<(0.03)*(0.97)/396>= -1.142

A bit low for my liking, but statistically it is within the significance level.

Set:

((26/396)-0.05)/sqrt<(0.05)*(0.95)/396>= 1.430

Still statistically ok.

Rare:

((34/396)-0.07)/sqrt<(0.07)*(0.93)/396>= 1.237

This is also ok.

Magical:

((328/396)-0.85)/sqrt<(0.85)*(0.15)/396>= -1.210

Darn it I almost feel cheated out of magical rings, but statistically this too is ok.

Thus the statistics say we have to reject the null hypothesis that Blizzard is lying to us on what our gambling outcome is. As a bit of trivia if I had only gotten 5 soj’s then Blizzard would be lying (6 and they would still be ok).

It’s been several years since stat class and so the point of this exercise was partly to exercise my brain and also to get those whiney people who say they “can’t” gamble for soj’s to just be quiet, only complain after they have a nice big n to back up their claims.

Yeem

Edit: spelling