Is Wine-Tasting Junk Science?
npr.org
npr.org
Really awesome video, thanks.
The true test of the validity of wine-tasting isn't whether experts can tell "good" wine from "bad" wine, but whether experts can reliably describe the same wine with the same language with an acceptably small margin of error.
My ideal experiment would give blindfolded wine experts a number of wines to taste. They would rate each wine using some appropriate standardized scale that would measure a number of relevant qualities of the wine. The experts would be given multiple samples of each wine, randomized appropriately. The experiment would measure how closely each expert mirrored their own rating of each wine on subsequent tastings.
This is basically a slightly more complex version of the famous (and hypothetical) "Lady tasting tea" experiment [1].
I like this setup because the ratings of each expert are compared only to their own previous ratings. I think this minimizes ambiguity and external sources of error.
My father-in-law Thomas is from India. He likes whiskey including Indian brands; his colleague was a Scotch whisky snob. At a party he challenged the snob to name all the whiskeys on the bar.
Thomas went into another room with the bottles in a bag; came out carrying a tray with half a dozen shot glasses. The snob tasted each one, make erudite remarks about their qualities and selected a brand for each one. Thomas opened the bag and showed that only the Indian bottle had the seal cracked.
Needless to say Thomas doesn't get invited over there any more.
The hard part is finding qualified experts who are willing to participate. The experiment would be most interesting and uncontroversial if it was done with experts whose credentials were unassailable.
There often are people who can discern these things but they're much fewer and further between than people realize. Of course, those in the first category who agree with my statements will never have their mind changed, they'll just assume that they're in the second category. The likelihood of that being true has a lot of zeros after the decimal.
The problem with articles like these is that many "experts" fall into the first category, not the second. This is not without merit though as often these are the same "experts" that people trust to make buying decisions.
But, I do think that Thomas' test was a bit unfair. He arranged as if it was a test of different whiskeys but then gave the same one every time. That isn't a double blind trial and it physiologically manipulates the participant. It makes a good story but I don't think it proves anything. Furthermore, you really need a larger sample size to say whether or not "snobs" can distinguish the real thing from the fake. If you are testing just one "snob" it would help to run multiple randomized trials.
There's an interesting documentary on some guys taking their master sommelier exam on Netflix called "Somm" who show an impressive array of these skills.
It's then everybody's job to taste again and see if they can put the glasses back in order. So far, nobody I've done this with has performed better than chance with vodka tastings and people barely perform better than chance with red wine tastings, even when they're 3 different varietals.
The test tends to come out after at least a few drinks though so most people's palates are at least somewhat compromised.
But that's the point of calling it "junk". What they notice (and thusly communicate) about the wine often simply isn't there. They certainly do "notice" things, but what they notice is often imaginary.
A good bottle of wine can certainly be enjoyed. It's a question of where they enjoyment stems from, and how much credence we put in the idea of some wines being significantly different and better than others.
In beer tasting, we talk about off-flavors like green apple, butterscotch, cooked cabbage, or flavors like spicy/peppery, cloves, bananas, citrus, passionfruit, etc. None of these are imaginary, as each of these are linked to a specific chemical compound, such as acetaldehyde, diacetyl, DMS, various phenols and esters, and of course hop oils and various humulones, cohumulones, etc. from the hops.
What would help in wine tasting would be to link each aroma and flavor to specific chemical compounds, it would make detection more objective and make it possible to tell apart the very real aromas and flavors from the imaginary ones.
The same thing has been done with other fields dealing with vague and loosely-defined concepts: stock-picking, sports commentary, etc.
The wine world is notoriously pretentious.
It's good to know locations, varietals and aging a little because it can help you narrow the window down to find a wine you'll probably like well enough vs ones you won't. I for example, tend not to like Malbecs. Why? I dunno, just something in the flavor. I also don't like Mountain Dew.
Wine can help accent or suppress flavors in your food as well. Just like drinking beer with pizza tastes good, you probably wouldn't drink coke with your morning cereal.
If you really want to get into wines, there seems to be a correlation between rating and flavor complexity. And there's all kinds of ways to taste (or miss) complexity in wines. Just like a good bbq sauce has elements of sweet, sour, spicy, etc., wine can be more or less complex in flavors. But not everybody likes complex flavors and that's fine. That's easier in fact since complexity is usually a result of more expensive wine making processes like aging so your local Two-Buck Chuck might make those people happy. But detecting the differences or preferences for flavor complexity is like detecting the differences between two different batches of your grandmother's home-made Spaghetti sauce. You can probably describe it (the first had more garlic, the second more basil), but absolute comparisons are really impossible and don't really matter all that much. It's like asking "which is better a sponge or the number 7?"
Something revealed to me during trips to various places where a glass of wine with lunch or dinner is normal, is that wine is just another drink. Some of the better meals I've had were had with a 2 euro 1/2 L carafe of the local watered down table wine. In these places, wine is just a drink, like juice, water or soda. People who drink wine with every meal really don't care what the wine is so long as it doesn't taste bad.
You can probably find more than enough wines anybody will like for under $10 a bottle.
It's just a drink.
I highly recommend the documentary Somm [0]. It's not an easy thing to become a master wine taster.
The best trained wine tasters can use deductive reasoning to narrow down the wine's identity, but a significant portion seems to be visual. Arguably, they are only effective at this because wine taxonomy is defined by visual characteristics. In a double blind taste test, I highly doubt it's possible to precisely identify the wine.
I like wine generally, but most of what I know of it comes from the documentary Somm, (on Netflix) - a sommelier's job is to tell a story - and to do this s/he must very knowledgeable and articulate about wine, its history, and its present. The history is basically a memory test, tasting wine is articulating what the drinker might experience in words, and having a palate honed enough to determine exactly what the wine is and where it's from.
All three, except the articulation, are verifiable quantitative tests. Even the during articulation the sommelier cannot stray too far from what is conventionally accepted.
I'm not harping too much on the sommelier service test, because the article is about wine tasting, but other sides apply. People train themselves to learn about wines in detail, and to differentiate wines right down to the highest level of classification and regionality just by just smell and taste alone.
I would also say getting to know an expert to the point where they know your own individual tastes really helps. Perhaps it is this aspect that people are getting hung up, everyone will have their own tastes, enter the expert who can categorize this for you, etc.
Wine recommended to you by trusted experts is not better than just blindly picking a bottle. Or at least, that is not something you can "say with certainty" if you haven't tested it. (And if you do test it, the results are likely to surprise you.) An awful lot of the experience of wine turns out to be the result of your expectations of it.
However its also been proven that people think that an expensive wine tastes better than a cheap one. So psychology and expectations play a big role.
I would advocate that we get mass spectrometers out there in smart phones so that we can scan the wine and tell its contents. Not only do we want to test the wine contents, we also want to scan our meat to check that it is what is says it is and we also want to scan vegetables for pesticide residue.
I had a couple of sleeves for wine bottles, so I decided to put them over the bottles and do a blind taste test. Of the 5 other people, 3 preferred the cheaper wine, 1 was neutral, and 1 preferred the more expensive one.
You don't need wine sleeves-- we've used paper bags, or you can decant in a double-blind safe way.
No one said science was easy, but someone's got to do it. ;-) You might even save yourself some money.
In fact there are several wine tasting training sets for sale, with tens of different wine aromas.
As humans OTOH, our smell is inferior to, say, dogs.
I think this is what people should take away from this article.
The "junk science" part is when self-described "experts" assign arbitrary percentile scores based on the tasting.
If you like it great.
This doesn't mean all aspects are purely taste, like: acidity, tanninity, type of grape, etc
since i've done this (in the past ten years) my enjoyment of wine has gone up.
For example, I tend to like wines with medium tannins and oak flavor, deep plum and tobacco notes. If there's spice flavors or not I'm also fine. I find I usually can find what I want in Chilean Cabs, and have less luck with other wines.
However, I've liked other kinds of wines too. Where I live the local Cabs are generally terrible, but the local Chambourcins, which have few of the qualities I generally look for, are good drinks too.