Email or username:

Password:

Forgot your password?
neatchee's posts Post Back to profile
Top-level
René M. Grabow
5 comments
neatchee

@yours_truly the black pixels simply indicate the most rare combinations. It's a little misleading because the scale goes something like 90th through 97th percentile: dark orange, 98th percentile: grey, 99th percentile: black. So while a bright yellow or white pixel next to an orange pixel carries meaning - here's a unique value that holds some significance to humans - a black pixel next to an orange pixel isn't very meaningful. It's usually just going to be the statistically rarest value along rare values, like a min/max function.

(1/2)

@yours_truly the black pixels simply indicate the most rare combinations. It's a little misleading because the scale goes something like 90th through 97th percentile: dark orange, 98th percentile: grey, 99th percentile: black. So while a bright yellow or white pixel next to an orange pixel carries meaning - here's a unique value that holds some significance to humans - a black pixel next to an orange pixel isn't very meaningful. It's usually just going to be the statistically rarest value along rare...

14 May at 10:11 | Open on urusai.social
neatchee

@yours_truly Now, if you can find a black inside an island of white/yellow pixels, THAT would be interesting. It would probably correlate to values tied to superstitions.

And upon consideration, I wouldn't be surprised if there are at least a few black pixels of that nature

(2/2)

14 May at 10:11 | Open on urusai.social
neatchee

@yours_truly 8068 having zero records is interesting but feels like an anomaly in the data, or a recording error?

Though maybe it's some kind of uncanny valley? Something like: it's just close enough to looking non-random that people looking for arbitrary values avoid it, while also being just barely arbitrary enough to never catch the attention of people who prefer patterns.

Pure speculation. Looking at the raw data where it suddenly jumps from a steady decrease in frequency down to zero for just the last value, I get the feeling it's just an error in the data.

Or maybe it's the author's debit pin that they "removed" from the data lol

@yours_truly 8068 having zero records is interesting but feels like an anomaly in the data, or a recording error?

Though maybe it's some kind of uncanny valley? Something like: it's just close enough to looking non-random that people looking for arbitrary values avoid it, while also being just barely arbitrary enough to never catch the attention of people who prefer patterns.

14 May at 10:21 | Open on urusai.social
Alain Dellepiane :eit: :pm:

@yours_truly @neatchee Seeing it a zero makes me think about a data error 😄

14 May at 11:20 | Open on localization.cafe
René M. Grabow

@aran @neatchee

Yes, maybe.
It could be due to the way the data was collected.

14 May at 12:01 | Open on troet.cafe
Go Up