Anonymous ID: 42db8e Dec. 15, 2018, 5:13 p.m. No.4327701   🗄️.is 🔗kun   >>7720 >>7752 >>7808 >>7815 >>8265 >>8279

>>4327556 (pb)

 

This Anon did this linear trend analysis. I simply went to Archives and grabbed the first post of every archive cohort (the 1st, 11th, and 21st of each month). I plotted first post number, date (I ignored time), and bread number. To fill in the post October gaps, I just grabbed two breads at random from the catalog. I was looking for gross trends, not granularity. Feel free to do the same. I didn't drill down to a bread-by-bread analysis because the gross linear trend was troubling (for the reasons you apprehend). Maybe CM or BO can share some metadata with us?

Anonymous ID: 42db8e Dec. 15, 2018, 5:32 p.m. No.4327901   🗄️.is 🔗kun

>>4327808

>https://qanon.news/Analytics/ScatterPlot

I've seen that, Anon, and it's good for what it is. What I was trying to get at is posting rate. Unless we're bandwidth limited, the rate should change as folks come and go, go to bed, go on holiday, give up because they think it's a LARP, etc. A bandwidth limit would imply that during busy times, posting is limited (which as far as I know only occurs during Q drop breads). THe longer trend shouldn't be linear with that sort of r-squared value.

Anonymous ID: 42db8e Dec. 15, 2018, 5:40 p.m. No.4328006   🗄️.is 🔗kun   >>8081 >>8277

>>4327937

OK, so on any social medium, posting board, etc., people come and people go. Maybe they lose interest. Maybe their internet gets cut off. Maybe they are newbs. Maybe they die. Whatever, the rate at which people post here should change over time. Very few posters per hour (or bread, or date, or whatever time metric) early on in the process, with more coming in later (growth and a faster rate) for lots of people giving up (shrinking and a slower rate). A linear rate suggests that either it's 1) you and me (and by "you" I mean that "you" might be a computer), or 2) there are so many people posting here that the comings and goings exactly cancel out. There are other possibilities, like someone if manipulating the post date/time to make this appear linear, but such linearity is suspect. It looks manufactured, not organic. That's all.

Anonymous ID: 42db8e Dec. 15, 2018, 5:48 p.m. No.4328118   🗄️.is 🔗kun   >>8137

>>4328081

Maybe. But a coordinated team of human posters seems unlikely to hit that r-squared value. That means that 99.6 percent of the variance in post rate is explained by the trend. since the beginning of this board…

 

The other 0.4 percent is residual (this is a crude explanation - go easy on me statfags)

Anonymous ID: 42db8e Dec. 15, 2018, 5:52 p.m. No.4328169   🗄️.is 🔗kun

>>4328137

 

I already did. This gets to the day to day granularity that I was talking about. But drop a linear trend line on those data and let's see the r-squared value. It's going to be linear. Spoopy linear…