User blog comment:Flonnaru/Results of Non-Drum Expeditions sparkle effect/@comment-24430017-20150609005754/@comment-25283969-20150609104457

in statistic you get the gaussian bellfunction and the results are somewhere there with different possibilities. simply by having RNG with you,you might get GS rates which are very spread around the estimated chance. if RNG is messing with you,your results simply go in a wrong direction,its not fault of the user,its fault of RNG : v. the point of eliminating is to make the confidence interval much smaller to make a preciser statement about the unknown GS rate while removing samples which are still correct but simply spread afar from the real GS rate. You perform this test correctly by writing down either great success or success after return of the expedition. I dont know how you expect people doing that wrong.