My crude understanding is, we don't have a good way to calibrate above-average scores, but it is a good measure for below average (provided there aren't arbitra

My crude understanding is, we don't have a good way to calibrate above-average scores, but it is a good measure for below average (provided there aren't arbitrary barriers like the language it's in).