5 ms·
From my understanding of the methodology, I don't understand what conclusions I could possibly draw from reading these reviews. My experience has been that peo
by ShabbyDoo 17y ago
From my understanding of the methodology, I don't understand what conclusions I could possibly draw from reading these reviews. My experience has been that people who don't drink much wine tend to prefer sweeter whites over those more dry. If you asked 500 of those people to compare a bunch of Chards in the $15/bottle range, they would end up choosing the sweetest of the bunch. As someone who drinks wine frequently, I would probably be upset if I paid $15 for those bottles. Like Netflix recommendations, I want to know which wines I will likely find enjoyable.
How can this problem be solved? I want to know which wines are liked by people who tend to like the same things I do. Could one do something like Netflix recommendations for wine? Would every user have to be calibrated by a bunch of blind tastings? Is there a more efficient proxy for calibration? Could a set of small, unlabeled bottles be sent to one's home where each represents a wine of a know, objective quality (oaked white, etc.)?
Unlike Netflix, I want to know if I'll like a $20 bottle more than another $15 one. It's not like movies are different prices, and we want to know if the $10 rental is sufficiently funnier than the $2 rental to justify the price differential.
Is there some way to crowdsource good wine data?
- compumike 17y agoHaven't looked for wine, but for beer the concept exists: http://beerriot.com/ http://beerriot.com/ Apparently I haven't rated enough beers to get recommendations yet... but perhaps if a bunch of HN people sign up and add their own likes/dislikes, it'll have enough data to get started. Also might be of interest that this is by one of the Basho (Riak / Webmachine / Erlang_JS / etc) developers (and a friend of mine): http://blog.beerriot.com/ http://blog.beerriot.com/
- andrewcooke 17y agothe comparison with netflix is a little misleading, i think, because there's a much clearer difference between films than their is between wines (famously, people have rated white wine + dye as red wine with no problems). pulling numbers out of my ass i would guess that grape variety, colour / tannin content, sugar concentration and acidity would cover all you needed if context wasn't so important. but context does dominate, so if you have faith in the site you'll enjoy whatever it says it is good... [actually, the more i think about this, the more i like the idea of something like you suggest based only on hard numbers for factors like i gave. if nothing else, it would be a lot of fun following different trails...]
- andrewcooke 17y agooops: and % alcohol.
- dminor 17y agoI believe cellartracker has community reviews of most wines, although I've never used it. There are also sites like corkd.com, etc.
- argv_empty 17y agoThere are a number of wine review sites out there, though I don't know of any that give you an easy way to figure out which other users' tastes match yours.