So I tracked 2,000 drops
The purpose was to get an idea of their model and figure out what was the best way to obtain the infinite amount of gear required. It wasn't really to find out what their exact "drop rate" is. If y...
Forum Discussion
"JoseIrimia;890707" wrote:
Thank you so much for the work and the explanations. Finally something with real numbers and reliable statistics.
I would like to ask about of one the points you made: You noticed quite long stretches of lower or higher drop rate. When having a probability like 1/3 or 1/5, such deviations should be extremely rare. Yet, they happen and often enough, but in the long run they correct themselves to the expected drop rate.
If you have a 6 side dice (similar to a purple drop rate), less 100 throws should be enough to determine whether is loaded or not with a confidence level of more than 95%.
I would like to ask to a statician or a programmer whether is possible that whether the probability for a particular fluctuates within a period of time, and in the very long run the integrated probability is the expected.
Actually, such deviations are more common than what humans perceive. It's why if you asked a computer and a human to write down a sequence of 100 heads/tails you'd easily be able to identify which is the computer and which is the human; humans don't expect the long stretches of luck. Another example is roulette, it isn't very uncommon to see a board with all black and all red, even though red and black have a 47.36% of happening (on a double 0 board).
Can you restate your last question, not entirely sure what you are asking there? Are you asking if a model can have fluctuating probabilities in time (.2 at one point in time and .5 at other times) if it can converge on a predetermined probability? If so, yes, you just have to schedule the durations to the appropriate time frames.
