UrbanDrinks uses a so-called bayesian rating system to rank places. This is simply a way of accounting for the number of times a place has been rated. For example, if place A has one 5-star rating, and place B has seven 4-star ratings, we should list place B ahead of place A because we have much more information and confidence in the rating.
We do this by adding five 3.5-star ratings to each place when doing the ranking calculation. We also require that a place has at least two ratings before it can show up in the rankings (i.e., the default 3.5 rating doesn't count until there are at least two real ratings). This number may change in the future.
Note: we only use this bayesian ranking method when showing the top-rated places, not when showing an individual place.
We will never pad our ratings in a way that inflates them or makes it so no place can have a rating below, say, three stars.
Please let us know if you think we can improve this system. Cheers!