Thoughts on the latest updates? Top 250, etc.

Couldn't disagree more- if you look at my prior post on the big movers on the lists the beers that fell down the list include super hype beers like Assassin, and 2 Cantillons! The beers that gained the most include some classics from New Glarus but there is High-westified which was brand new this past 6 months, and Mexican Cake is on the list too.

Overall, the list now has a bit more overlap with the classics from the top 250 of fame, but there are still a ton of new beers on the list - just a lot fewer really hyped ones...

 

Seriously? you think 150 characters is verbose? It is 2 sentences. Your post above is 402 characters - so almost 3 times more than required for your review to count...

I think the new system is great, and all it asks is you spend a minute sharing a few words about what you experienced trying a particular beer.

 

LOVE LOVE LOVE LOVE LOVE LOVE LOVE LOVE LOVE all the new changes, I know I've gotten into so many shouting matches against the bros and to see ratings be important again makes me so so so happy

 

You guys, is NO ONE concerned about what this may have done to the bottom 250 beers?!

 

Nope - it seems about right

 

I have only been around these parts less than a year, but I love the changes too - frankly if you can't be bothered to share a few words and put some thought into a review I couldn't give two shits about your rating.

My only concern is that people are going to go back in and add slipshod reviews to get their ratings to count again - adding a bunch of useless noise when I try and filter to find actual thoughtful reviews.

I made a copy of the top 250 (have been doing some analysis on the overlap to the old list) and I am going to take a snap shot every few weeks for a while to see how it morphs. I am expecting things to be pretty dynamic for the next few months as stuff gets reviews added, and things move up and down.

 

As a former PA resident, I feel your pain about case laws. It's actually that same thought process I have in the trading world but I see it the opposite way. I want to know if the beer I'm trading and shipping I'm going to pay is worth this bottle I'm getting.

To me what is more useful, one or two in depth reviews by random people (probably friends drinking from the same bottle) who may or may not know what they are talking about, or 100 anonymous ratings, giving you a statistical sense on whether the beer is good or bad? Now you won't know why its good or bad, but with enough people they can be more useful then a couple reviews. Now this is an extreme case, and would prefer to see 10, 20 or more reviews to get an understanding, but sometimes a bunch of ratings can be more useful then 1 or 2 reviews. Making this change will just decrease the amount of data points without really making a lot more reviews (IMO). Add to that, all the beers added in the last 2 or so years are screwed, because people didn't put in reviews and likely don't have a good enough memory or notes to add one now.

 

I much prefer using actual reviews again over hads when scoring. This should not really be thought of as a new thing, but going back to how people originally reviewed and expressed their opinions on this site. The way I see it, the Aps, quick "hads" or short one liners is relatively new with the Twitter and Untappd sites. I originally chose BA over RB because the content was so much better. Part of that was people actually reviewing beers and places, rather than just saying yay or na with no real content behind those scores. I personally get nothing from that.

 

Continuing with the analysis - breweries that got pwned hard and lost the most beers from the list.
To make sure I was complete and fair, I also added back any beers they gained so it is net change:



Interestingly it looks like a who's who of hyped and traded beers

 

I appreciate it. The art of a review has fallen by the wayside. More reviews = better understanding of beer and better information spread overall. I always review my beers; it is worth the extra time. Good move guys!

 

And to me, this is very sad. The beers I've had from these breweries have raised the bar to what a good beer is. These breweries are forcing the industry to get better and produce more enjoyable tasting beers. Fremont's coffee KDS blew away the standard (BCBCS) causing me to rethink how good coffee BA stouts can get. SARA Saison Bernice was the first beer to truly be sour, tart, fruity, juicy all in one to make an enjoyable fruit beer I didn't think was possible. Trillium, Tree House, NEBCO, and Hill Farmstead keep redefining how much taste and drinkability can be put into an IPA. De Garde and Side Project are revolutionizing what Berliner's and American Wilds can do. They are taking what Russian River did, and making it better, something I didn't think possible a few years ago. The greatness that these breweries are doing should be showcased and emphasized. Anyone who likes one of those styles, needs to know about these breweries so they can try and enjoy their beers, either by trades, friends, beer-cation's or however. The loss of these beers from the top 250 show the list went from the not really debated best list of beers around to just another top beer list that clearly is missing the great beers being made (like every other internet list).

I would love for everyone to go in and add reviews so these beer can be brought back to the exposure they deserve, but after 2+ years of minimal reviews and not everyone having good memory/notes, it's not going to happen. So instead, we get a list showing all the classic beers everyone already knew about that pale in comparison to what was lost.

 

Interestingly of the beers that I have had that have been kicked from the list, I was disappointed in all but one of them - they almost all failed to live up to expectations, and were frankly not worth the effort I had to put in to get them.

Personally, I disagree with you, and I love the new system - to me the hyped and over-rated beers have dropped from the list. I guess we can agree to disagree on that though.

 

Another stat this time based on the ratings themselves:

The average rating (yes, the scaled bayesian one) in the old list was 4.45, and 69 beers had over a 4.5
In the new list, the average is 4.32, and only 11 beers have over a 4.5

Looks like those people who write reviews tend to have less upward skew in their ratings overall, and there are much fewer '100' beers now.

 

Sounds reasonable and what I'd sort of expected. Thanks for doing some serious stats on this.

 

This whole discussion is really very disappointing, with people just voicing anecdotal evidence about their personal favorite beers and their overall average score, and how this raised or lowered them. Is everyone missing the real point here about throwing out the majority of ratings data just because the person didn't write a 150 character "review", which for some arbitrarily biased reason means their rating is more meaningful? Does anyone in this forum really understand what statistics are and that better, more reliable and meaningful statistical results are achieved when you have a really large sample size? A few "bad" quick reviews that rate a beer 1 or 5 are meaningless when you have hundreds or thousands of ratings in total. A few "bad" full reviews count a lot more when there are only a handful of them! Who doesn't understand this basic statistical principal?

 

I like the change, but my suggestion would be to give "hads" some weight, maybe 25% of a review? It is useful data and will fix the problem of only old beers showing up on the list now.It will make it harder for new breweries to crack the list but not nearly impossible like it is now with this change. A nice middle ground.


Also, can you please please pretty please consider making a "whale" or "highly rated rare beer" list that includes every beer (archived ones too), and doesn't have a minimum # of reviews, but uses the same algorithm? That would be awesome, especially for people who like seeking out the big boys.

 

Not counting ratings would work well if it was always that way. Having all the thoughtful ratings I've entered over the last couple years mean nothing unless I add a review is crap. So much useful data tossed to the curb.
Yes, the new system will probably get more reviews written but it will probably lose 75% of the users moving forward. I for one am not going to write reviews or take notes when I'm out sharing bottles with friends and thats usually where new stuff gets popped. The ease of use is the only reason I do it. Change things for the better not just for the sake of change.

Westy 12 in the number 3 position? Thats just shows the list is in the past and it's going to take years before its relevant again.

The single most important part of this website has been tossed in the garbage. Huge fail in my book.
At least we still have untapped.

 

I look at the ratings system against the numbers personally, 4.5 is 100, so anything above that truly has to blow me away. Also the way I look at reviewing vs ratings, I generally will see a hyped beer really much higher if its just a tasting; I only review full sized beers and that usually pulls down my opinion(of the high rated ones). Opposite effect on the low rated ones. I will think less of something but a full review usually outputs a higher overall opinion(there are exceptions to both sides though). Also, cool thing I noticed with most of the local newer breweries that had previously lower overall beer portfolio scores: Their overall portfolio beer avg went up right around where I felt about them personally as a whole, even though before it was a fair amount lower. I feel like a lot of the hardcore reviewers rate to style most of the time, while there were probably a good portion of raters that put up low numbers because they don't like the style etc. Just my personal assumption.

 

I presume you have read the thread. Perhaps you noticed the summary stats that I have been posting (I am a degree in stats and do data analytics and market research for a living). As a statistician I am noticing a much more normal distribution of scores, with way less skew than before. For the beer types with lots of beers (like IPA and APA) the medians now line up much more closely with the means. This is because the ratings without reviews have an upward skew - people who don't review also tend to rate higher than those who do - even when there are huge sample sizes (like with Pliny etc...)

I can't help but note that you have 2,500 ratings, zero reviews and a grand total of 4 posts ever on this website in over 5 years. All 4 of those posts are on this thread. While I appreciate you joining in, have you never felt the urge to share with your fellow beer advocates what you actually thought about any of those 2,500 beers?

It is interesting that the people who have been unhappy on this thread also tend to be the people with enormous numbers of beers ticked off for the time they have been on this site - seems like there are 2 kinds of users and the pendulum has swung back toward the beer advocates and away from the tickers.

 

Welcome! I have something wrong with me; I actually enjoy data stuff