Wine-tasting: it's junk science | Hacker News Reader

Wine-tasting: it's junk science

theguardian.com

17 pointsmiles10 comments

> Some of the judges were far worse, others better – with around one in 10 varying their scores by just plus or minus two... A wine deemed to be a good 90 would be rated as an acceptable 86 by the same judge minutes later and then an excellent 94.

Plus or minus 5/100 isn't bad at all for anything requiring human judgement. Its certainly within what I'd expect ratings to swing for judged olympic competitions like figure skating.

... plus, if you actually set up the rating process the right way, things would probably look even better.

A lot of these studies set up strawman investigations of the process, and then use the results to claim that that people can't discriminate meaningfully between wines, which is a huge leap and an unwarranted conclusion.

I got curious after reading your comment, and it appears that even though the max rating in wine scoring is 100, the min value is 50. So it's /50, really... still relatively accurate(?)

https://en.m.wikipedia.org/wiki/Wine_rating

Not really, when there is only 8 points between "acceptable" and "excellent". In this case, 5 points here and there makes a huge difference.

Perhaps that's a problem with how the scores are interpreted. Imo they should ideally not have that many grades, perhaps only 1-5 or 1-10. Or if they do, they should be a 92 +/- 5

Like I mentioned, judged olympic competitions typically come out pretty close as well, and people argue all the time about the judge's decisions too. But its not as bad because the scale runs from 1-10.

Another example would be essay grading, though I don't know exactly how much that deviates. The SAT essay section even has guidelines on how to rate essays and what to do if the two human grader's ratings disagree too much.

Anything judged by humans is going to be subject to this kind of thing.

Surprise surprise, taste is subjective.

The byline is misleading, wine experts most certainly can distinguish wines by grape, region, and even vintage with remarkable accuracy, they just don't grade them consistently.

Here's my tip for drinking good wine on the cheap. Buy the Trader Joe's label. Not two buck chuck, but the ones called "Trader Joe's Reserve", "VINTJS", and the like (they have several variations).

I buy a bottle when they have a new one (usually in the flyer they send out), usually for $8-15, and if I like it, buy a case or so. Better hit rate for me than Wine Spectator scores or whatever.

Not sure how they source the wines, but the consistency is pretty good.

It's simply entertainment and business, just like the Oscars. Does anyone seriously think otherwise? And would they care? People seem to take the "oscars" seriously.

For me, 'expert' wine recommendations are rarely useful. But, crowdsourced reviews have been remarkably consistent. A vintage with a mean score of 4.1 of 5 (over, say, 100 reviews) is almost always better than a wine with a mean score of 3.4. But, it's still subjective. For instance, I like Zinfandels and Petite Sirahs, so reviews of Pinot Noirs are less useful to me.

hey don't knock it. sure its subjective. but a nice wine is a rare pleasure.

(2013)

Reply on news.ycombinator.com