Quote Originally Posted by Kazemon15 View Post
Please make me understand why ignoring 80% of the playerbase/population is a good statistic for things? And if that is how other companies do so as well, basing it on 20%, then yes, I will say that. Because it's largely ignoring the majority and should be looked at.
You are not 'ignoring 80% of the population, math is neat like that. Let's say I was organizing some FFXIV event in which 100 people showed up and I wanted to ask who among the attendees like Wuk Lamat, but asking all 100 of them is impractical so I ask 15 of them at random(important) and 7 said no and 3 said yes(no response for the rest, so unsure). There are two possibilities here, either a lot more people dislike Wuk Lamat, or I got really unlucky and just happened to pick Wuk Lamat haters. We can mathematically measure exactly how unlucky I'd have to be. In the given example, there is a 90% chance my sample group result is within 20% of the full population, meaning between 50% and 90% of the people dislike Wuk Lamat. 10% of the time I'd get unlucky and the real number falls out of that range. The bigger the sample, the more accurate the results, and so long as the sample is truly random, the size of the total population is largely irrelevant. A random sample of 3000 people (assuming truly random picking and that they are truthful) is enough to predict the entire world population's views within 3% regardless of how big the population is. This is a pretty standard fact in statistics, it's just the central limit theorem applied to iid Bernoulli variables and plugging in 2sigma then solving for n with confidence interval. You don't need to flip a 2 sided coin a million times to get a prediction on results any more than you need to poll every FFXIV player about DT. I can copy paste my AP statistics book formulae if you're interested in the actual math, but there's plenty of online calculators that can solve almost all statistical problems for you.