This is totally unrelated, but I was just reading about Stanford's OL, and our center is named Khalil Wilkes, or KWilk. Seeing you reply to this thread made me double-take.
I literally came onto BT to ask an untappd question about ratings. Found this on Untappd thanks to your thread.
Go to bold at bottom if TLDR.
JulianB because you do this.
$avg_num_ratings // Average number of ratings in all beers (split for commercial vs homebrew) $avg_rating // Average rating for all beers (split for commercial vs home-brew) $this_num_rating // Number of votes for the beer in question $this_rating // Pure Average Rating for the beer in Question
$bayesian_rating = ( ($avg_num_ratings * $avg_rating) + ($this_num_rating * this_rating) ) / ($avg_num_ratings + $this_num_rating);
I'll rephrase it
Ntot // Average number of ratings in all beers (split for commercial vs homebrew)
Ra // Average rating for all beers (split for commercial vs home-brew)
N // Number of votes for the beer in question
R // Pure Average Rating for the beer in Question
$bayesian_rating = ( (Ntot* Ra) + (N * R) ) / (Ntot + N);
So even if a beer has ten 4.5 reviews (average) it'd come in at:
((100*3.5)+(10*4.5))/(100+10) = 3.59
And I'm seeing almost literally the following for Real Ale Benedictum (in the formula).
((40*3.5)+(6*4.2))/(40+6) = 3.59
One could say this implies the average number of ratings per beer is around 40.
Anyway, what I really came on here for was to make sure that check-ins without ratings don't hurt a beer. They don't seem to but a confirmation would be great. Any input is appreciated.
Cheers!
Tangential majority of post brought to you by a bomber of barleywine Intihared and me loosing weight despite at least 500 Cal/week of beer on top of foods.