A experiment for all - Area lvl 85 runs

I had nothing better to do at work today then think of Diablo, so I came up with a semi-great idea. What about getting as many as possible people from the forum to make a experiment on Alvl 85 dropps? The choice of char to mf with is up to the person joining the experiment, and so is the MF%. And we can make a discussion here on what areas to run. My choices would be CS, WSK, The Pits and Ancient Tunnles. How long this schould take is altso up for discussion here.

So all I need is some feedback for this idea :smiley: Are you in?|||I would say per person hundred runs, this is really doable with a good (good not godly) equiped character, with ten people helping it would already be 1000 runs. I will join in as soon as I get my new hammer going

EDIT: The number of players in the game should also be factored in, as well as other benefits (Getting XP yourself, helping others level etc etc)|||The SPF does this all the time... you can look at result charts if you search for them.|||Naturally, I vote for 400+:smiley:|||The more the better, but then, it is not my time! |||I vote 400+. May good drops come to you. |||i vote for 300, i think its enough anyway )|||If you just want to have some information about who drops how much of what, better use ATMA, except the reason to do this at all is to proove ATMA or to proove it wrong.

I think you need at least 100,000 runs in total to make meaningful statistics about e.g. all uniques which might be considered as statistical evidence that ATMA is correct or wrong and your statistics about really rare ones like Tyrael's or unique, class-specific TC87 armor still won't be very accurate. For that purpose, better consider making a few millions of visits in total.

To be honest, IMO it's a big waste of time. As already said, have a look at the single player forum, that's where the masochist MFers are.|||krischan, are you aware that precision usually goes with the square-root of the number of measurements?

i.e. 100 000 runs isn't 100 better than 1000 runs, but only 10 times as good?|||Well, I studied physics which involved a lot of statistics.

With that logic you can as well say "Why doing 100 runs when it's just 10 times better than a single one ?". No matter if it pays off linearly or not, it's the only way to add precision.

It doesn't depend on the number of events, but on the number of events with the result you are looking for, so when finding let's say one Tyrael's in 100,000 runs, that's an average of 10 in 1,000,000 runs with a standard deviation of sqrt(10)~=3.16. That means if you do several series of a million runs each, the average difference between the expected value and the actual value for each series of 1,000,000 runs has an average of 3.16. If it isn't within certain limits, chances are good that the theory which predicted the value is wrong.