Just read another thread mentioning skewed album scores. One might see an album with a score of 30, which might give the impression that it's not a very good album, but if you look at the reviews you might see one 0 and two 90s. To me that looks like an album that is highly thought of, except for one hater, rather than one that's low quality.

So when computing album scores, throw out the outliers. You can toss the highest & lowest scores like the Olympics, or you can calculate a true statistical outlier and toss those. That would be better, actually. It would require a certain minimum number of reviews before any score is shown, but that might be a good idea anyway. The text of the review can still be shown.

Throwing out outliers would also limit score wars, in which albums will receive a dozen or more reviews (Iron Maiden) serving little purpose other than to affect the score, with fanboys giving everything 100% while non-fanboys give scores in the 50s-60s to intentionally balance out high scores that they don't think are "correct".

What do you think, sirs?

Why do I have a feeling of déjà-vu...?

(0 + 90 + 90)/3 = 60...