/vt/ - Virtual Youtubers

>>66494549
marine's vod views are absurdly good, if that's what you're interested in
If you ignore all the weird extra messing around this guy does and just look at the raw vod views per minute streamed in the greyed out column you can see that Marine is far and away the most successful streamer by that metric, or at least she has been for the last half-year or so

I was looking at this views/time thing and there seems to me to be a lot of really weird methodological choices that introduce a lot of noise and weird biasing.

One of the biggest ones though from my perspective is that the ranking for each is converted to a fraction that gets averaged out with your ranking in other time series. The issue the way I see it is that longer stream durations have much smaller denominators because fewer streamers have e.g. 7 hour stream vods. So the result is that your rank 1 at 7~8 hours gets converted into 1/38 or something wheras your rank 1 at 1~2 hours gets converted into 1/400. One of those numbers is like 10x the size of the other, you're basically just getting penalized for bothering to have streamed in more stream duration zones? What's the point of that?

There are plenty of other issues but like... in general this seems really quite silly to me. How has this been going so long without any methodological fixes?

I was looking at this views/time shit and I don't understand the reasoning behind the weighted average of the ranking represented as a fraction

Doesn't that just reward high-ranking short-duration streamers since the denominator for the short-duration streams is so much larger, meaning they get a lower fraction? Why would we want that to influence the rankings...??