Earlier this year I removed the minimum and adjusted the algorithm, which allowed us to experiment with ranking and scoring all eligible (active) beers. After gathering feedback and reflecting on the impact, I've decided to revert back to requiring 10 or more ratings for a beer to receive a BeerAdvocate Score and be ranked. To be more specific, this applies to the algorithmic minimum used to calculate a weighted average for each beer, which is then used to generate BA Score as well as ranking a beer against all beers and beers within its style (used for our top rated lists). Anyway. While it was good to try all of this, I learned (once again) that I shouldn't change things that work (our rating system) in an attempt to address an underlying issue (arguably too many beers and not enough users to rate them all) by trying to force a desired outcome. That said, I'll continue to explore other solutions and listen to your feedback. tl;dr: the algorithm (and min) has basically been restored to what is was before March 3 and everything has been updated. Thanks!
I thought the changes were an improvement and thus think this is a step back. It makes the lists a lot less useful IMO, outside of the main ones (BA250, Fame) and larger areas (California, Illinois, Belgium, etc.) The Bayesian minimum of 10 always seemed way too small to me from a formula standpoint. Also, I thought it was fine for a beer in a category like grodziskie with only 6 check-ins to be on the top list for that category.
