[ageitgey]

It's a test measuring outlier knowledge. By definition, most of the words will be words that never come up in common usage (and thus words of limited utility outside of reading classic literature). You shouldn't feel bad if you don't know 90% of them.

[gethly]

it's still stupid. proper way to measure vocabulary is to divide words into groups by usage(it's easy to analyse articles or videos from last 10 years to get real and not theoretical data), from most used to least used. you end up, for example, with 10 groups with simple words like "then" and end up with rare words, like "discombobulate". you pick 10 words from each group and ask the user about knowledge of these words. then you can build a solid overview of the user's understanding or knowledge of the vocabulary.

[AndrewDucker]

It'll be doing something like that. Ask you a known word. Ask you a less known word. Get more and more unlikely until it can place where your knowledge falls off.

So once it's verified that you know a couple of semi-common words it won't ask you any more common words at all.