Hrm, pardon the quote tags. They're so tacky... but:
Carsetius wrote:
Grep: I highly doubt that it is anything that complex
WiccanDiva wrote:
I presume the only part of the Count you are talking about it the end number giving the daily high count.
We as players can't really use count over a 24 hour period, nor do we know when the samples are actually taken. I decided to make a rather ideal model of the playercount in order to find the average in terms of a variable we
could reliably get as players, the maximum player count in a day (or peak.)
Code:
24
/\
| Peak-1 ( pi * hour ) (Peak-1 )
| ------ sin ( --------- - pi) + (------ + 1) d(hour)
| 2 ( 12 ) ( 2 )
\/
hour=1
25 < ____________________________________________________________
24
This creates a curve representing a very idealistic daily player count fluctuation, finds the area under this curve in a 24-hour period, and then divides it by 24 for a very accurate, analog version of a mean player count for that day.
Basically, the mean of a sine wave is half of its range, or whatever, which is easy enough, but we wanted to understand the mean in terms of the amplitude (or peak) as a variable instead.
Under this ideal situation, the peak we see when we type "count" in-game should be over twice our target mean.
Depending on whether this ideal model represents a higher or lower player activity than we can readily expect, our solution would be either above or below our goal, of course. I suspect, but cannot really confirm, that this ideal model is a higher estimate of count than truly occurs, perhaps with weekends as an exception on account of the longer and potentially multiple climaxes.