Scoring is very counter-intuitive.

[alexyans]

Players appear to rank higher even though the 1-norm and 2-norm vector distances do not justify that. Sometimes even the human perception of the colors they chose appears to be further from the target color than those of other players.

[ssimono]

So far the calculation is made on Euclidian distance corrected with some factors as found here: https://en.wikipedia.org/wiki/Color_difference?oldformat=true

There have been many attempts to weight RGB values to better fit human perception, where the components are commonly weighted (red 30%, green 59%, and blue 11%), however these are demonstratively worse at color determinations and are properly the contributions to the brightness of these colors, rather than to the degree to which human vision has less tolerance for these colors. The closer approximations would be more properly coefficients of 2, 4, and 3

I understand that this can lead to some frustration, I am thinking about using the CIE metrics created for this reason

[alexyans]

That sounds great 👍

[ssimono]

Just found on HackerNews a library that provides a color space said to match human perception https://www.kuon.ch/post/2020-03-08-hsluv/

Interesting topic, I am leaving this for myself in case I (or someone) wants to implement it in this project