Some of the colours on the screen don\'t look like they are in the pots, at least on my screen.
Don\'t worry too much about the on screen display - that doesn\'t actually tell you much of anything unless you have calibrated your monitor to the same output that I have calibrated the scanning equipment to. If you want to do that though, it is calibrated to FujiChrome IT8.
I suspect it works by comparing HTML hex codes for colours...
Nope, I use sRGB as the starting point. The actual comparisons are performed using CIE Lab color space formulas for mixed media formats. It is effectively the same method used at the big paint companies (Sherwin Williams, Valspar, Dutch Boy...) in order to handle paint matching at the factories as well as the machines that paint stores use to match color samples that people bring in.
Also, comparing GW to P3, why hasn\'t sanguin base or highlight been mentioned as a comparison to the GW reds?
There are a lot of different variables that will effect the end paint sample used. If you have a few that you suspect are wrong, let me know. The FAQ does explain things a bit more.
When I have ones that I know need to be corrected, I order up three of the pots from different suppliers (hopefully getting paint from different batches in the process) and take new samples averaging those three together.
Anywho, if you have any problems with the colors - let me know. I can\'t fix what I don\'t know is broken.