IIR Seems To Increase the Drop Rates of the Wisdom Scrolls - PoE

Posted by Richard Betts on May 4th, 2018

I'm still adding more data to it, but the data is starting to become quite meaningful. For those that don't know, if P is <0.05% that means something almost definitely happened, there's only a 0.05% chance that it happened by fluke.

The primary take away from this is the following:

IIR Seems to increase the drop rates of the following:

• Wisdom Scrolls
• Orb of Chance
• Chromatic Orb
• Orb of Fusing (Probably)

And seems to lower the drop rates of the following:

• Armourer's Scrap
• Blacksmith's Scrap
• Orb of Transumation (Probably, but only by a small amount)
• Orb of Alchemy (Possibly, wut?)
• Jeweller's Orb (Possibly)
• The rest I don't have nearly enough data on to speak for.

However, some basic questions about your assumptions.
You are referring to a p-value based on an alpha level of 0.05 correct? so your alpha error is 5% not, 0.05%, thus your chance that this result is significant but actually not significantly different would be 1/20. At least that it what I get when i quickly glimpse over your first formula in the spread sheet.

Also, why is your sample size so different between your two treatments. It makes it less comparable and might violate assumptions of your statistical tests.

Also you have 42 replications of your control group. no offense but with such large variability and the amount of degress of freedom, I don't see how that is even remotely robust for the conclusions that you have drawn preliminarly.

Finally, you are conducting a lot of statistical tests on the same data set and thus might be prone to alpha-error-inflation, a simple and conservative approach would be to adjust your p-value via the Bonferroni method. Simply divide your p-value threshold by the number of statistical tests conducted. Do you want to poe items for trade? If you do, you can visit U4GM to have more guides and methods about this.