2011-09-15, 02:20 AM
RADRaze2KX Wrote:Why would I need to prove this within 1% accuracy? I can get 500 EE scrolls if necessary and mass produce 500 of an item to scroll, with say i guess a staff, 700 100% scrolls to use on all of them. If after the first 100, the probability doesn't seem to be near 80%, (say 50% pass/fail) why continue? Double it, say the trial works on 70% in 200 trials, at what point do you stop and say "the rate truly is 80%" ? If 4 out of 5 work, that's truly 80% ... if 400 out of 500 work, that's 80% .. what happens when the probability is different per user? What if I fail 40 out of 50 and you pass 45 out of 50? Which one of us gets to determine the pass/fail rate of the scroll as seen by the community?
Because that's the number of trials you need to do to be reasonably sure. You ask at what point do you stop and say "the rate truly is 80%?" That's the point where you stop.
If you did 5 and 4 works and you obtain the probability of 80%, with a 46% maximum error then you can't say for sure can you?
100 trials will get you within 10%. 500 trials will get you within 5%. (with 99% certainty)

