>>8257456Ah, I see. In that case, the lower medians make sense since I think only a few Holos can get a stream average of 10k, because it takes time to ramp up and usually there's a cliff when they start reading SC.
The problem I have with using stream average is that many JP holos spend significant time reading SCs in the latter half and it absolutely tanks the average of the stream. That skews the stream average to be less representative than the peaks imo. By the time they read SC, the stream is effectively "over" so I don't think they should be punished for going extra.
Peaks are usually deceptive for stuff like costumes, but those wouldn't matter much in median calcs anyways since they're in the far end. Compared to average (which includes SC dragging it down a lot), I think peaks are a better representation of the stream.
For example, in this stream, the bulk of the stream (MC) was above 10k, but she spent an extra 1.5 hrs on SCs and it destroyed the average of the stream.
https://holo.poi.cat/stream/HiyMsaD-DfsAlso, for the sake of uniformity, there's another anon here that keeps track of NijiEN's monthly avg and median, and he uses the peaks of each stream. So comparing Holo and Niji numbers would be a pain in the butt if one is using peaks and the other is using average.
Anyways, just my 2 cents. Would appreciate it if you could do a peaks median too.
