(26 Dec 12, 12:54PM)SKB Wrote:Even with normalization by #map loads, you can't reasonably compare 1078 loads for the 3rd place with 13 loads for the 4th place. You have made no determination of the # of map loads needed for a representative sample. I don't remember much stats, but iirc you could probably calculate a #of samples needed to achieve some confidence in the average?(25 Dec 12, 11:47PM)Ronald_Reagan Wrote: When some maps get loaded ~10 times and other maps 9000 times, you can't expect to have accurate results.
Technically yes, the bigger the sample the better. It doesn't really matter, though, as all NET values are "normalised" by map loads. Originally, the limit was at least 5 map loads, but seeing DM urban having NET value of 4.8 made me reconsider the sample size.
Also, I haven't thought about this too much, but normalizing for time-of-day might also clarify the results. My vague idea is that different maps may not be evenly distributed across time => different times w/ different #s of players and different timezones of players online may have different leaving/joining behaviors.
*Edit: I like the current results for mode. #map load normalization doesn't matter so much there, so those are interesting right now.