This is like one of those eye tests where they switch between lenses narrowing in on the prescription. The question is really what shades of turquoise is more blue than green.

I got 80 which is close enough, I think it’s really only the extremes that are meaningful. I tried simply alternating green-blue etc and got 60. I think adding some randomness and taking more samples (more questions) would help - I was worried that the prior color left a residual effect as a relative comparison was easier than absolute comparison. The extra random samples could help give an idea of confidence in that middle zone.