... And the forum ate my reply, damnit.
Long story short, Most programming languages have stats packages that will easily crunch this data for you, and can automate pretty much the whole process. You just want to give them the drop records you experienced (If you kill three skeletons and they drop 0, 1, and 2 bone chips you'd give them "0,1,2") and let them give you the mean (expected outcome from one kill) and standard deviation. 95% confidence of where the actual mean resides is within 2 standard deviations either way, so Mean +/- 2*stddev. Alternatively StdDev is simple to calculate, just look up wiki and use the sample formula, not population.
You won't have any issues with items that drop over '100%' as you're describing the expected outcome in truth, not the chance of getting 1 of the item, so just treat it like anything else and people will understand that a 160%+/-10% drop rate means they should expect 1.6 silks per kill and that you're fairly confident the real average is somewhere between 150-170%.
|