How are beers recommended?
The recommendation of beers is done via two heuristics. In the first heuristic, we base ourselves on similarities between the beers. For instance, the category they belong to, alcohol percentage, fermentation methods, etc. If certain beers are recommended to you via this heuristic, it means they are similar to one or more beers in your beer lists.
For the second heuristic, we base ourselves on your own beer taste (derived from your beer lists) and the beer tastes of other users. The underlying concept is simple. For instance, say that most users, when they give a high score to beer A, also give a high score to beer B. If you gave a high score to beer A, we can infer that you will probably also like beer B. As more users make this link between beer A and B, it becomes more likely that beer B will be recommended to you.
Both heuristics are applied, and the beers that best correspond to your beer taste are recommended. The reason a certain beer was recommended to you is displayed below the recommendation. Beers that already occur in one of your beer lists (tasted or to taste) are automatically filtered from the results. Note that the effectiveness of the second heuristic strongly depends on the number of users; the recommendation feature will improve in accuracy as more users sign up and create beer lists.