Something of interest that I noticed is the fairly constant gap between areas of low colour acuity
I tried it without much effort on my so-so Dell desktop display and got 59. Tried again on my RMBP with even less effort and got 14. However, I did have trouble in the same places both times so it's certainly demonstrating something.
Dell: http://imgur.com/MFvcSky
RMBP: http://imgur.com/61WlVQS
Green/Red street lights are easy to tell apart, but the amber light in the middle can look very close to either depending on the light level outside. A red flower amongst a green leaves is something I would overlook, unless it was pointed out to me. Then I would have to search for the flower. While looking directly at it, its pretty apparent its different from the green leaves but to spot it without previous knowledge that its there is a difficult task (This situation came up yesterday while someone was pointing out a wild rose among a hedge.).
Overall it I would say it doesn't have an effect on my daily life. I can still wire up a RJ-45 connector under good light conditions (I use my phones flashlight to do it). Oddly enough, being colourblind has made me better than my coworkers at wiring up Cat6, as I double and triple check each of my connectors quite thoroughly to make sure there is no mistakes (As me making a mistake will directly be blamed on colourblindness and my abilities will be called into question, if a coworker makes a mistake it is just that - a mistake).
Anyone else here take the XKCD color survey back in 2010: http://blog.xkcd.com/2010/05/03/color-survey-results/
I was presented with an awful lot of colors I didn't have a better word for than "tan", "beige", "green", etc, so I expected that my "color acuity score" would be poor.
Looks like I was having trouble without something to compare to, and having trouble naming, rather than having trouble identifying differences between shares. Good to know!
...I think something is wrong with their best/worst scores, actually. About 80% of the age/gender brackets say the best is 0 and the worst is 1520, and that consistency is weird to begin with, but some have really dramatic outliers. Women aged 50-59 range from negative 162 to 410,378,090!
a) your capacity to stare at a screen without tearing up, blurring, etc.
and
b) the quality and color profile both native and gamut-corrected etc. of your monitor?
I swear this is much more a test of my current monitor settings vs. ambient light conditions, than anything serious about my own color perception.
A physical version with color chips and the option to work under a variety of lights (LED, halogen, CF, incandescent, sunlight...) at whim would be way more revealing.
But yeah, harder to code in JS.
My results: http://screencast.com/t/s5ohpINp
Also, by the end of the process I felt like I was going to be sick, not sure if it was the brightness (my eyes are extremely sensitive to light and I normally keep the brightness at 1 bar but I had it at full brightness for this).
EDIT: Based on a brain damage study with 48 patients and 48 healthy controls "a total error score between 20 and 100 was taken as the range of normal competence for discrimination."[1]
[1] http://www.perceptionweb.com/abstract.cgi?id=v990416, "Scoring efficiency on the Farnsworth - Munsell 100-Hue test after brain damage"
I have fun running different sorting algorithms on the tiles. The trick to getting a perfect score is doing a pass where you switch every pair of adjacent tiles. This will double the color delta on the edges if the tile is correct and halve the delta if the position was wrong. I always catch at least a blue or two on this pass.
A part of it is the fact this is on a computer monitor though. If you only got a few wrong, it's probably not a big deal.
I could tell a properly-placed color apart from the colors two or three spaces away from it, but not usually from its immediate neighbors. My main strategy was to assume that, if a color looked identical to each of its neighbors, it was in the right place. I got a 16, which seems decent enough.
I'm red-green colorblind, and got a 90. Ouch. It looked pretty good to me!
I got a perfect score but am neither artistic nor fashionable.
Using a TN LCD panel with an effective 6 bits per color channel and a compressed dynamic range would be the opposite of cheating, it would be playing with two hands tied behind your back, so to speak.
Using a wide-gamut IPS panel that is properly calibrated would be more like "cheating" (though not actually cheating if your own color perception is really what you want to test, as in that case the monitor should reproduce color as accurately as possible).
