It should probably alternate between blue/notblue... green/notgreen. I hit the same wall. Second question asked if blue/green when it was neither... and I really mean neither. I don't see cyan as a shade of blue or green, rather much like I don't see green as a shade of blue or yellow.

There's no way for me to answer truthfully whether teal is blue or green. It is neither. Anything I give gives a false answer. The data is invalid.

But reproducibility should be the point. As a result of the structure it approaches an asymptote from one side or the other. I took it once and approached from green and my greenness was 77%, a second time it approached from blue and my blueness was 68%.

A test that allows an answer of neither would deliver more information (transition points and an error bar) without failing to identify a distribution in the population taking the test.

>correctness is not the point. binary choice is the whole point. because my blue may not be your blue...

Realistically there is a broad range that we all can acknowledge is neither, but is instead teal, and forcing a binary choice means people just choose randomly.

It's so remarkable how many people here refuse to understand your point. It's like, there is no right or wrong, no perfect score, just pure subjectiveness, and they can't handle it. If I wasn't convinced this site is entirely bots before, I might be now....