Average visit time isn’t a very useful stat. Is your average visit time 15 minutes because 50% of players play 30 minutes and the other 50% play 1 minute? Is it 15 minutes because there’s a tight grouping around 15 minutes? Averages don’t tell you very much. We need better stats for visit time. An example of what would be useful is something @TheAmazeman posted on Twitter a while back
https://twitter.com/TheAmazemanRBLX/status/712054088171528192
If each time-point was visualized as a spectrum (like brighter color in vertical dimension, proportional to how many players played that long), that would be good.
Adding standard deviation to the dev stats would solve this, as you can very quickly calculate that 68% of your players play between -1 and +1 standard dev of the mean.
This would definitely help. However, average and standard deviation alone do a poor job of representing data that is irregular. Those metrics are good at helping understand the variation of a data set but do a poor job of showing where that variation is.
Yeah, thats why I think some kind of per-day spectrum or such would be good.
Like divide each day (or whatever the interval is), to N segments on the Y axis (1 minute, 2 minutes, 3 minutes…), and for each count how many visits fall in that segment. Then just display their relative values in graph, for each point in time.
I don’t mean to discredit this suggestion, but maybe someone(s) should make a script which you can drop into your game and it takes statistics on simple stuff like this.
Also, @Waterlimon, I think the word you’re looking for is histogram.
Kind of like a 3D histogram, I guess? (time, visit length, count)