Forgot your password?
typodupeerror
Science Technology

Chemists Build App That Could Identify Cheap Replacements For Luxury Wines 206

Posted by samzenpus
from the tastes-great dept.
schliz writes "Australian startup Wine Cue is combining the chemical composition of wines with customer ratings for what it hopes to be a more objective wine recommendation engine than existing systems that are based on historical transactions. The technology is likely to reach the market as a smartphone app, and could be used to identify cheap alternatives to expensive bottles."
This discussion has been archived. No new comments can be posted.

Chemists Build App That Could Identify Cheap Replacements For Luxury Wines

Comments Filter:
  • by icebike (68054) on Sunday June 09, 2013 @11:01PM (#43957623)

    If there is one thing that needs more objectivity its wine tasting.
    Too often the results are the opinion of the person who bought the bottle, and too seldom is there truly blind taste testing by people not already familiar with the vintage.

    • by Nerdfest (867930) on Sunday June 09, 2013 @11:15PM (#43957675)

      Even with blind tastings there is subjectivity, as people sense of taste and smell is quite varied. I tend to be good at picking up secondary aromas (not the primary fruit) but amd not as good at picking out some of the subtle fruit smells. It all come down to chemicals ... esters and other compounds, that can be measured objectively, but for now is still quite expensive to do accurately. Any good sommelier can generally pick out a cheaper example of an expensive wine for you based on what you like though. It may not be *as* good as the expensive one, but it is a game of diminishing returns for the most part, although it is occasionally possible to get a *better* wine for less money. Wine sells for what the market will bear, based on origin, availability, and reputation.

      • by Mr. Freeman (933986) on Sunday June 09, 2013 @11:26PM (#43957745)
        You're also probably not as good at picking out anything as you think you are. If you've never done any kind of double-blind testing to find out then your assumptions are likely nothing more than conformation bias. Same thing with "experienced sommeliers"

        Example: You drink a wine and say that you taste X. The next person at the table hears this and therefore tastes X. It also probably works in reverse. That's not to say that you couldn't tell the difference between two wines that are drastically different, but subtle differences are likely imperceptible.
        • by LordLimecat (1103839) on Monday June 10, 2013 @02:37AM (#43958459)

          Does it make any sense to speak of confirmation bias and objectivity when talking about "taste"?

          • it does if you're trying to indirectly identify chemicals in a glass of wine.

          • It does indeed make sense, and there are many who are striving to construct an objective set of criteria for evaluating wines.

            However, even this can be foiled by expectations. There have been several instances where researchers have found that supposedly knowledgeable judges tasting wines in a blind (as in blindfolded) context have been shown to be unable to distinguish white wines from red.

            Objectively, that might stand to reason, given that both are likely to carry a similar array of organic compound
          • by Hatta (162192)

            Yes. If you slap a higher price tag on a cheap bottle of wine it "tastes" better. There is no price receptor on your tongue, so there must be something else at work. That's confirmation bias.

        • by CastrTroy (595695)
          What really makes it interesting is that different people can taste different things. In highschool while studying genetics we learned about a chemical that some people perceived as bitter, while other's didn't taste it at all (probably this one [utah.edu]). So it's completely probable that a wine that one person might think tastes terrible is actually quite pleasant to others. Even if a wine doesn't have any bitter compounds, it's not unlikely that somebody like a wine taster might have a heightened sense of taste
      • by Maxo-Texas (864189) on Monday June 10, 2013 @01:09AM (#43958233)

        Why on earth would people who eat different foods and have different taste profiles and come from different ethnic heritages be expected to like wines equally.

        I've had several blind tastings.

        For most people, the ability to taste a difference tops out in the $20-$30 a bottle range.
        I've only known one person who had the ability to finely discriminate wine and he came from the new york area.

        At one tasting- the one bottle he disliked, everyone else liked.

        There is a tremendous difference at the lower end because many of the less expensive wines are either

        a) Just bad (and just about anyone can tell this)
        b) or they are "Thin" (watered down, one note) which anyone can taste pretty quickly and easily in comparison to a good wine.

        But there are plenty of wines good enough for 14-18 a bottle.
        And plenty of wines that are good enough after you are tipsy for $9-$14 a bottle.

        The truly great wines require an experienced and truly great wine tasting ability.

        And why give truly great wine to people who can't tell the difference anyway (i.e. most of us).

        • Why on earth would people who eat different foods and have different taste profiles and come from different ethnic heritages be expected to like wines equally.

          ...

          There is a tremendous difference at the lower end because many of the less expensive wines are either

          a) Just bad (and just about anyone can tell this)
          b) or they are "Thin" (watered down, one note) which anyone can taste pretty quickly and easily in comparison to a good wine.

          But there are plenty of wines good enough for 14-18 a bottle.

          ...

          And why give truly great wine to people who can't tell the difference anyway (i.e. most of us).

          Raises hand. That would be me. I don't know how to buy wine in the first place except to know that some wines are sweeter than others and there are reds vs. whites and the pink ones that fall in between.

          Of course, that means I don't have to spend a lot of money on wine in order to be happy with my purchase. I suppose I do better selecting beer which I consume several times in a month. If this app gives me some way to evaluate wine in the first place it could be a win for me. However I'm a little skeptical t

      • by Ihlosi (895663)
        Even with blind tastings there is subjectivity, as people sense of taste and smell is quite varied.

        Afaik, this was a trap soft drink manufacturers fell into. They had taste tests with dozens or even hundreds of participants. Later on, it was found out that human genetics alone can produce thousands of different "taste types", depending on whether a persons smell/taste receptors can pick up the presence of certain chemical compounds.

        Hence, having a hundred people in a taste test is far from being a repres

    • by mjwx (966435)

      If there is one thing that needs more objectivity its wine tasting.
      Too often the results are the opinion of the person who bought the bottle, and too seldom is there truly blind taste testing by people not already familiar with the vintage.

      So they should invent a device that can detect a pack of Winnie Red's (Cheap but powerful cigarettes) and recommend a box of 4 penny dark (cheap red wine) as they wont taste anything anyway?

    • by ultranova (717540)

      If there is one thing that needs more objectivity its wine tasting.

      But taste is by definition completely subjective. It's like trying to make an algorithm to detect good art: at the absolute best you can simply predict how a given test person will judge, since there is no objective quality to measure.

      Besides, wine is an excellent luxury item: a $1000 bottle doesn't really take any more resources to make than a $0.99 bottle, despite resulting in great perceived difference in lifestyle, so from society's poi

      • By "taste" do you mean the sense of taste? Some things are objectively salty, sweet, acidic, etc.

        I can't speak to $1000 bottles of wine, but that $100 bottle took more resources on average than the $0.99 bottle. The land those grapes grew on is expensive.

  • by Jeremy Erwin (2054) on Sunday June 09, 2013 @11:05PM (#43957641) Journal

    Black Stump Bourdeaux is rightly praised as a peppermint flavored burgundy, whilst a good Sydney Syrup can rank with any of the world's best sugary wines.

    • I have a terrible palate but I've tasted these notes and it was kinda cool.

      There are probably a dozen that anyone can taste after drinking a couple dozen bottles of a particular type.

      I've had a "peppery" wine and a "chocolatey" wine- which was bizarre because as you say, they were just grapes.

      I've also had 3 perfect pairings and 1 near miss. When that happens- it's like magic. Each sip of wine makes the food taste better and each bite of the food makes the wine taste better in a swirling dance of gustator

      • I have to admit-- I did try to find out whether peppermint was a standard wine aroma, and instead came across this choice tidbit on wikipedia [wikipedia.org]

        The German Wine Institute has created a special German language version of the Aroma Wheel meant to be specially adapted to German wines, with one wheel for white wines and one wheel for red wines.[4] However, in the translation they removed the petroleum smell (and the entire "chemical" category) from the white wine wheel, despite the fact that mature Riesling wines

      • by drinkypoo (153816)

        They're not just grapes, though. Besides the yeast pitched, there's other stuff in the grapes. don't have a link but there was an article on how if your grapes don't contain wasp spit they won't make great wine...

  • If I were to do this stuff myself, I would probably use Partial Least Squares and build a regression model using the chemical composition as X, and the customer ratings as Y. Or depending on the number of variables in the chemical composition compared to the number of samples (wines), one of the Sparse Partial Least Squares variants might prove to give better predictions (and it would also be interesting to see which variables in X it discards as less important).

    So, any word on what they do?
    • by Xest (935314)

      No idea what they do but I did exactly this for a university project using neural networks. It was some years ago so I can't really remember the details of what exact data we had, I just remember having a dataset of chemical composition data for a good number of wines as a training set.

      At the end of the day this is just a straightforward classification problem so any number of statistical classification methods should work just fine.

  • Bum Wine (Score:4, Insightful)

    by deadhammer (576762) on Sunday June 09, 2013 @11:37PM (#43957803)

    BumWine.com [bumwine.com] lists the only wines you'll ever need.

    • by Nimey (114278)

      "Tales of Cisco-induced semi-psychotic fits are common. Often, people on a Cisco binge end up curled into a fetal ball, shuddering and muttering paranoid rants. Nudity and violence may well be involved too."

      Truer words were never spoken.

      • by cellocgw (617879)

        Tales of Cisco-induced semi-psychotic fits are common. Often, people on a Cisco binge end up curled into a fetal ball, shuddering and muttering paranoid rants. Nudity and violence may well be involved too.

        Wait a sec, Cisco wine does that, too?

  • For most people the price of wine is not about taste. It is about exclusivity and following the rules of the pack. You pay this price for this bottle of wine because that is what your peer group is doing. Otherwise why would it be so easy to forge wine labels and sell then as the real thing [nytimes.com]? If you have connections and providence people don't seem to know the difference.

    That is not to say that expensive wine does not provide value. You are paying for vintage grapes and expert winemakers, which all cos

    • This is why, once you've convinced the hipsters that the $11 wine bottle tastes like the $300 wine bottle, the $11 wine bottle will be a $45 wine bottle.
  • by reverseengineer (580922) on Sunday June 09, 2013 @11:43PM (#43957837)

    Welcome to Wine Cue!

    INPUT: Chateau Petrus, 1998 vintage, Pomerol primarily of Merlot grapes, estimated retail 3500USD

    RECOMMENDATION: Charles Shaw, 2010 vintage, Merlot, estimated retail 2USD

    • From testing and experience, the $3 wines are undrinkable unless ice cold and nasty then.
      $6-$12 is fine for drinking with food or getting tipsy but thin.

      At $13 + you start getting decent wines.
      Most folks taste buds seem to top out at $25 bucks a bottle.

      And you shouldn't be wasting a $60+ bottle of wine to get tipsy or if you can't tell the difference or you just don't like it. I can tell the difference a bit but don't interpret it as better.

      • by drinkypoo (153816)

        You can often find some place near you that has cheap wine, too. The guy who owns our local Grocery Outlet also bought out a liquor distributor. He sells good wines cheap in the Grocery Outlet. Too bad about the beer selection, which is mostly fake craft beers, but I do enjoy buying amaretto and schnapps for six bucks a bottle. I don't drink it — that's what beer is for — but it's awesome for cooking. You could spend five bucks for a tiny bottle of extract...

        Point is, I don't even drink wine (re

      • Try holding some blind tastings. I have. Among casual wine drinkers, there's no correlation between the price of the bottle and the prefernce for what's inside. At my own tasting parties (where I do blind taste testings), the Trader Joes Coastal Zinfandel continues to crush the competition for $6/bottle. Despite what people claim to like, clean, fruity zinfandels are reliable winners.

        I've started growing grapes and making my own wines, so I'm looking forward to see how my own vintage fares at my next wi

      • by neminem (561346)

        We (me and my fiancee) actually really, really like the Fresh & Easy brand 2 dollar wines, especially the white (which is actually two dollars. When Trader Joe's upped the prices on their 2 dollar wine last year, our local F&E actually put up a slightly snarky sign that explicitly was not naming any names, but you knew who they were referring to, saying hey, our wine is still 2 dollars.)

  • Target house brand box red wine. That's right, you buy it at Target (at locations where they're allowed to sell wine).

    The three varieties, Merlot, Shiraz, and blend are all good. It's like the best $12 bottle you've ever had -- not a typical $12 bottle, the best. The box is $16 and contains the equivalent of four bottles, of course with the self-sealing spigot and collapsing plastic bladder to prevent oxidation. Stays fresh for weeks or even months after opening -- provides a glass a day for three weeks.

  • by Jeremy Erwin (2054) on Monday June 10, 2013 @12:06AM (#43957953) Journal

    I've had some spectacular wines. No, no, not the wines that cost hundreds of euros per bottle. but wines that could be described as "WOW. I didn't know wine could do that". It would be nice to have an app that would suggest similar wines, based on a chemical spectrum instead of "that estate had a truly extraordinary summer, and more recent vintages have not faired as well."

    If a particular chemical is playing around with my brain,I want to know about it and be able to invite it around again sometime.

  • It's not just the basic chemicals but the molecules and how they are 'folded' which makes a MASSIVE difference to what happens.

    I predict this will be technically correct but completely useless, as seen in that classic joke about mathematicians:

    Two Physicists were riding in a hot air balloon and were blown off course sailing over a mountain trail, and were completely lost.

    They spotted a jogger running on the trail and they shouted "Can you tell us where we are?" After a few minutes, the jogger yelled
  • Of course you cannot tell how good a wine tastes by some chemical analysis.

    • There's a researcher (who is also a trained sommelier) who is using chemistry to produce unconventional wine/food pairings based on underlying shared aromatic compounds. Seems to be working well for him.

  • I used to live about 15 minutes drive from the Barossa Valley. You can definately taste the difference between a $50 bottle and a $10 bottle, but having said that I don't believe the $600-2000 bottles are justified at all. I highly recommend the Yalumba Signature and Octavius wines. Bought at the cellar door they're ~$50/bottle tastes amazing. Much better value than the Grange Hermitage people love to harp on about from Penfolds. That starts at $600/bottle and goes up from there. I've tasted both of those w
    • by tjb (226873)

      I haven't had the Grange Hermitage, but I have had a couple of bottles of the 1996 Penfolds 707, and even in 2010 it still tasted really young. I suspect the Grange may just need to be cellared for a *really* long time.

  • I am asking because beer companies discovered that their most avid customers couldn't taste the difference between their products and their competitors products.
    • by mjwx (966435)

      I am asking because beer companies discovered that their most avid customers couldn't taste the difference between their products and their competitors products.

      Crap beer tastes similar to other crap beer?

      I'm shocked at that revelation. Truly shocked.

      Here's another grand revelation for you... People who buy mainstream beer buy it because it's cheap, not because it tastes good.

      I'm willing to bet the beers in that test were 1. US mainstream beers. 2. Lagers. First off, the rest of the beer drinking world refers to #1 as "sex in a canoe" because it tastes "fucking close to water"* and as for #2 Lagers are designed to have no taste. Now comparing a semi-decent

  • Fuck me. Truly a first world problem.

  • There's an interesting interview with a wine writer named Jim Clarke [google.fr] in the book Cooking For Geeks.

    This guy says the four main variables in wine pairings are :

    • - acidity
    • - sweetness
    • - alcohol level
    • - astringency or tannins (for red wine)

    And I think it's refreshing, and also very sensible to think about wine in such "basic" terms. Even if you can detect all kinds of interesting flavours in wine, like world-class sommeliers do, I think those four variables are definitely going to influence your experience a l

    • Wouldn't it be nice if, in addition to alcoholic content, the labels on wine bottles clearly displayed the amount of sweetness, acidity, and astringency ? I'm talking about real numbers with some kind of scale.

      Yes, but misleading.

      We tried this at a wineyard. The guideline we were given was to stay away from anything but "dry" wine. The measurable amount of sugar left is used to distinguish between dry and sweet wines, but a good dry wine can have fruity aromas that will lead to a perceived sweetness, even though the fermentation has been completed. (and hardly any sugar is left)

  • I drink maybe a bottle of wine per year (a glass here, a glass there) plus maybe the equivalent of 2-3 bottles of sparkling wine (champagne, prosecco, or asti). Wine gives me awful hangovers, worse than overdrinking whisky or beer, and generally it doesn't appeal to me -- kind of bitter and unsatisfying.

    Until about 2 years ago if you had asked me about wine, my instincts would have been that it's 95% bullshit and 5% reality (the difference between jug table wine and a $20 bottle of wine).

    And then I got dra

  • I'd buy one of these if it could tell me if what I'm buying is what I think I'm buying and not some counterfeit...

    Would be nice if it worked for liquor as well given how much poison is out there on the market at this point.

  • than any other commodity, except maybe audio equipment. Anything that cuts through some of the bullshit is welcome.

    Of course, those of you with sophisticated palates who enjoy fine wines will have no use for such a mechanistic means of judgment and will disregard it. However, this development should please you as it provides yet another reason to turn your noses up at the unwashed masses who would be so ignorant as to select a wine based on chemical composition.

"It's curtains for you, Mighty Mouse! This gun is so futuristic that even *I* don't know how it works!" -- from Ralph Bakshi's Mighty Mouse

Working...