On MyAnimeList, while viewing another user's profile, the sidebar displays your affinity with them:

Affinity can be positive or negative, and (supposedly) represents how similar your scores are, and by extension - your tastes.

The affinity system has become a sort of gatekeeper, deciding whether two people on MAL should be friends, or whether one should trust recommendations from another person.

If you're interested in the math, the affinity calculation uses Pearson's correlation coefficient - you can read more at Wikipedia or the MAL thread that proposed a better affinity system.

Here are some thoughts from the lead dev for MAL's thoughts regarding affinity.

🔗 What does 100% affinity look like?

The difference between two users' scores is much more important than the scores itself. As an example, these two lists have maximum (+100%) affinity:

The above example represents a few problems with affinity. It doesn't require users' scores to be similar, only that the difference in scores be similar.

This might be useful, as it disregards a users scoring system; if someone rates all their shows around 7/10, and another rates their shows around 5/10, affinity will account for that.

However, it also means that you can get false positives. As an extreme example, this is also 100% affinity:

... even if these users have completely opposite tastes.

This is not to say affinity is a totally broken system - if two users have a similar rating system, affinity will be fairly accurate in representing how similar their (difference in) scores is.

The table below represents two users' lists and their scores, try changing Lulu's scores to increase their affinity.

🔗 Change Lulu's and Ginko scores to see how the affinity changes!

	Lulu's Score	Ginko's Score	Score Difference
			0
			+1
			+3
			-2
			-1
			-2