2014-11-14, 10:23 PM
actually we are closer to a solution than we thought
based on drop rate extrapolated from straatlamp and zapta's tests.
problem we have faced is the variation in the drops
so I have calculated the number of runs you need to do to overcome it, to basically have it such that the result is clearly one or the other
based on a 95% confidence interval. I have calculated how many tests you need to do so that the 95% CI do not overlap. when that happens the result is clearly one or the other.
maybe do a few more than the listed min number to be sure
bascially straat can do it in one run or two runs since he has so many equips getting 200%, so it produces a large enough difference
| tests | additive | mult | add expected drop | add 1.96 | mult expected rop | mult 1.96 | diff in means | sum of CI 95 range | min tests | ||
| fam and pendant | 1.6 | 1.2 | 1.8 | 1.92 | 72.35 | 16.11 | 77.17 | 16.60 | 4.82 | 32.72 | 46.01 |
| card/event pendant | 2 | 1.2 | 2.2 | 2.4 | 88.42 | 17.67 | 96.46 | 18.39 | 8.04 | 36.06 | 20.12 |
| card/event fam | 2 | 1.6 | 2.6 | 3.2 | 104.50 | 19.06 | 128.62 | 20.89 | 24.12 | 39.95 | 2.74 |
| straat pendant fam | 3 | 1.6 | 3.6 | 4.8 | 144.69 | 21.97 | 192.93 | 24.72 | 48.23 | 46.69 | 0.94 |
| straat pendant card | 3 | 2 | 4 | 6 | 160.77 | 22.96 | 241.16 | 26.89 | 80.39 | 49.86 | 0.38 |
based on drop rate extrapolated from straatlamp and zapta's tests.
problem we have faced is the variation in the drops
so I have calculated the number of runs you need to do to overcome it, to basically have it such that the result is clearly one or the other
based on a 95% confidence interval. I have calculated how many tests you need to do so that the 95% CI do not overlap. when that happens the result is clearly one or the other.
maybe do a few more than the listed min number to be sure
bascially straat can do it in one run or two runs since he has so many equips getting 200%, so it produces a large enough difference

