Rankings (based on "mean" rating data)

And I would say the freshness factor may add some credibility to that sentiment. Especially for beers that have to travel days for distro.

 

My point is that the mean is off for a state like Wyoming due to inflated beer ratings from biased BAs living in Wyoming. I think I'm safe to assume that hardly anyone living more than 300 miles from the state of Wyoming has tried a Wyoming Beer. How many beers have you tried from Wyoming? Probably zero unless you go skiing in Jackson Hole. How many different beers have you tried from California? Probably 50 or more. Therefore, California Beers have many, many more ratings from unbiased out of state BAs that give a more accurate rating of how good the beer actually is.

Chances are based on your example, you would be disappointed with the Wyoming Beer because on average what should be a 3.95 beer (the mean) is actually an overinflated 3.25. The California Beer on the other hand, with an average of 3.79 actually will be a 3.79.

 

Just told the wife. Forget that trip to California we were planning. We are going beer hunting in Wyoming

 

Careful, some beers with 50 ratings have fair ratings. 50 LOCAL ratings may be biased, but if you get a good sample of 50 reviewers from all over the country or the world, it's fine. A good example of that, I think, is http://www.beeradvocate.com/beer/profile/25314/199189/ Ras Soft Citra IPA - Omnipollo. 4.2 is about right, I feel it should even be 4.3. And there's only 11 reviews, 8 from the US, 1 from Romania, 1 from the UK and myself from France. I think it's a good representation of what the beer is. It's brewed in Belgium by a Swede, so 0 local has tasted it.

 

The margin of error with a sample size of 11 is huge. Feeling like the score is fair or accurate is irrelevant. That's the whole purpose of statistics.

 

Yeah but when hesitating whether or not buying a new beer, are you gonna wait months or years before the sample size is big enough ? No, you're gonna assess if the 10 people who already reviewed it can be trusted or are overly enthusiastic with no real perspective on it. Usually when all the reviewers find the same tasting notes (ie fruits, roastiness etc...) and pretty much gave the same grade, it's ok. If you see a couple of 5's or 4.8's real early with other 3.8s, beware.

 

Was she disappointed? (Nothing against Wyoming as a great place to live or visit as long as beer isn't involved.)

 

This list is hilarious. (The data, not the work by @bubseymour )
2 beers in the UK and Ireland that score above a 4.25 and so on.

Well, lots of people, and that's the issue. Perhaps many BeerAdvocate site users can agree on this notion, but those are the same people whose data created the list. Would someone in Bamberg be able to argue against it? Which brings me to:

Forgive me for being nerdy on this, but the only thing I'd argue against is the interpretation of this data. You're interpretation is using language like: "best of the best," great, decent, really bad, crappy... beer. This data is more of a measurement of BeerAdvocate user preferences than a measurement of quality or even general likability. Yeah, these can be very related, but let's not jump to conclusions of causality. So... in the example above, you'll have greater chances of grabbing a beer that many other BeerAdvocate users like. Apologies for nitpicking, but I think it's an important distinction even if most folks will roll their eyes at me.

 

I would love to see this without the weighted BA averages. JKs MVB is a 4.52, but rounds down to 4.3ish. Harsh. Same with those casey cuts 4.49=>4.14!!!