Blind Tasting

Home > Other > Blind Tasting > Page 9
Blind Tasting Page 9

by A. C. Houston


  "You've got to drink something besides wine sometimes," Rob laughs.

  The training station of cups has morphed into an an ink blotter. Using a different Q-Tip for each extraction, Cory places a tiny dab of wine from each of the four Dark Moon zinfandel vintages onto the blotter, four inches apart.

  Dawn is on the floor with one leg tucked under her, and the other with the recovering ankle, stretched out in front of her. She leans over the coffee table and marks a number by each drop to correspond to its bottle.

  Rob, crouched near her, places a drop of wine from the third bottle of Dark Moon onto a small piece torn from the blotter. This will be Cory's control for the next experiment.

  Snoots watches the trio with a mixture of curiosity and concern. They are occupying his area by the coffee table and they are pursuing an activity he's never seen before. He's not sure it's okay.

  He approaches Dawn and touches his nose to her hand, the one that is labeling dots on the blotter. He makes eye contact as if to inform her, whatever you are up to, I'm in charge and supervising it.

  Rob stands up and offers Dawn his hand. She takes it and he pulls her up, holding on just a moment longer than necessary. "How's the ankle?"

  "Much better, thanks to you." She gives him a quick, shy smile.

  Cory calls Snoots over and the dog reasserts his rightful place in front of the training station. He's ready for his task. Cory holds out the control and the dog looks at it with momentary interest, but doesn't sniff it. He looks back at Cory expectantly, where is my cup?

  "Maybe put the control blotter in a cup?" Dawn suggests.

  "Good idea." He takes a clean cup and puts the small control blotter inside, offering this to Snoots.

  The dog sniffs it and then hovers over the blotter with the four samples, he nostrils barely quivering. He points to the third spot, Dark Moon's Enchanted Hill label, zinfandel grapes from their higher altitude block. Exactly what Cory has for his control.

  "Dawn, you are brilliant!"

  "Any brilliance we're seeing belongs to Snootsy," she remarks in quiet exultation.

  Rob shakes his head in awe. "Man, at this point, that dog is a better judge of wine than I'll ever be."

  "I've been really curious to know what he'll do if the control isn't an exact match. Up until now, he's always been given a control that identically matches a wine in the sample set." There's quiet excitement in Cory's voice.

  Dawn ponders this. "The null hypothesis says he doesn't pick any of them."

  Cory's eyes meet hers. "But if he does."

  "If he does, my first reaction would be that he's just choosing something to get a treat."

  "Well, he didn't pick any of the Dark Moon zinfandels when I used cups. That supports the idea that he's trying to match some quality that he can identify."

  "Why don't you try really different wines. Different, at least by human standards, and see what he does," Dawn suggests.

  Cory puts down fresh blotter paper and dabs a drop from each of four already-opened wine bottles: a merlot, a sauvignon blanc, a pinot noir and a zinfandel. He puts a drop of chardonnay on a fresh control blotter.

  Snoots samples the control, briefly sniffs the offerings and looks away.

  "No match. Good," Dawn remarks.

  "I'd be back at square one if he had matched any of those," Cory tells her. "I believe he's honing in on similar characteristics in the wine that humans can also detect. He just does it a hell of lot faster and far more consistently."

  "What does he do when the wines are all, say zinfandel, but from different years?" Dawn asks.

  Cory goes to a whiteboard propped against the wall in his living room. He picks up a pen and begins to write on the whiteboard.

  "So, if the grape varietals are all different, he doesn't match them unless a sample and the control are from the same bottle of wine. In other words, control A matches sample A, but control E doesn't match samples A, B, C or D in this situation."

  A

  A b c d

  diff. varietals

  E

  a b c d

  diff. varietals

  Cory continues, "And, he easily matches a wine to another sample of itself in the following tests: single grape varietal all from one year, but from different wineries and, even more impressive, single grape varietal all from one year produced by just one winery, but from different vineyards."

  Cory adds these two scenarios to the whiteboard.

  C

  a b C d

  same varietal

  same vintage

  diff. wineries

  D

  a b c D

  same varietal

  same vintage

  same winery

  diff. vineyard blocks

  "So, now we're going to test one grape varietal from one year but where the control is not identical to any of the test samples." Cory writes a question mark for now.

  E (not identical to a test sample)

  ?

  a b c d

  same varietal

  same vintage

  diff. wineries

  He opens a bottle of 2007 Rockpile zinfandel from Grouse Winery and then begins to open his bottle of 2007 Todd French zinfandel that Dawn bought for him at The Sage's Cask.

  "Maybe you don't want to use that one?" she suggests.

  "These are all the zins I've got right now." He grins at her. "And anyway, can you think of a better time to open it?"

  He places a dab of the Todd French zinfandel onto a new control blotter and a drop of the different '07 Dark Moons and the '07 Grouse onto the test blotter, assigning each drop its ID.

  "Isn't your control a zinfandel blend?" Dawn asks, remembering that the tasting had featured zinfandel proprietary cuvées.

  "Yeah." Cory nods. "Another parameter to take into account. I really don't think he's going to match this cuvée to any of the samples, but we'll try it first, then switch the controls."

  He offers Snoots the control and the dog sniffs it, then the sample set. Without deliberation, he points to the drop from Grouse Winery.

  "He matched it," Dawn says, shaking her head in disappointment.

  Without a word, Cory fills two cups, one with the Todd French, the other with the Grouse. He tries both of them, then emits a soft, low whistle and offers the cups to Dawn.

  She carefully savors a mouthful from each. "They seem close. I'm no expert, but they seem damn close. My god, Cory."

  Cory rewards Snoots with a peanut butter dog biscuit and takes another sip from each cup. "I'll bet money that Todd French sourced the zinfandel in this blend from Grouse's Rockpile vines. In fact, it's probably mostly those Rockpiile grapes. Not so much of a blend, after all. But who has heard of Grouse Winery? A handful of foodies from San Francisco? It's a little place out on Dry Creek. It sells for about thirty bucks in the few restaurants in Sonoma that list it. On the other hand, Todd French 2007 Olympia is one of Pillar's Picks. 95 points."

  Rob laughs. "I guess Snoots just put Grouse on the map."

  Cory switches the control to one of the Dark Moon zinfandels, but this time Snoots doesn't pick any of the wines.

  Cory returns to his whiteboard and edits the last sketch.

  E (not identical, but close match!)

  a b C d

  same vintage

  diff. wineries

  shared vineyard block?

  "The source of the grapes seems to be a strong indicator," Dawn remarks. "It trumped any differences in the process of the winemaking that took place at Grouse versus Todd French."

  "Or maybe Grouse actually has a winemaker who can kick ass like Todd French," Cory replies.

  He looks at Dawn and Rob, shaking his head. "With six weeks of training Snoots can make these calls in just seconds. Without depending on the color, viscosity or mouth of the wine. He tastes, well smells, totally blind."

  "It's simply incredible." Dawn stares at the setup on the coffee table.

  Rob picks up the bottle of Todd French
zinfandel. "What if Snoots hadn't picked this?"

  Cory looks at him keenly. "It would mean there were no interesting similarities." He cups his hands under his chin. "I'm not sure what he's picking up on exactly to identify the wines. He's got possibly a million times more power to throw at the problem than we do, right, Dawn? And he knows what the task is and that was the huge breakthrough. Find the wine that matches the control, either identically or now, closely. I've wondered whether I should run his matches through a mass spectrometer to see what the chemical relationships actually are." He glances at Dawn. "But where do I get a mass spectrometer?"

  She shakes her head. "Forget about Genetica. We have a big one, but it's in use full time." She smiles at him. "Besides, we're a Level Two lab and I don't want to add to Greg's headaches by screwing around with unauthorized usage of equipment."

  Cory nods, still thinking. "I have a couple of friends in the physics department at Stanford."

  "Cory, I'm not sure what a mass spectrometry of the wine samples would tell you anyway in this case. It's a good technique for comparing the chemical properties of two substances, but two molecular shapes could be very similar yet result in very different scents for Snoots. Or for us, for that matter."

  Rob listens in fascination as she continues.

  "Take limonene, an essential citrus oil, a hydrocarbon. C10-H16. It has two isomers, the molecules are mirror images of each other, quite similar. But, one smells like lemon, the other like diesel oil. Or take hydrogen sulfide, H2S, and basic borane, BH3. Very different shapes, but almost identical scents. Add the tremendous complexities introduced by fermentation and the analysis becomes truly formidable."

  "I wonder." Cory becomes lost in thought for a moment, then looks at Dawn intently. "Today, Leonard Pillar is the word on wine. Consumers, wine shops and wineries around the world pay attention to his numbers, his points. Why? Because, like the guy at Sage's Cask said, his judgment is deeply consistent. His palate can repeatedly distinguish and identify many important qualities of wines. That's the value of a great critic."

  He looks from Dawn to Rob. "And now, Snoots has been trained to identify these types of similarities and qualities between wines. Only he can do it really fast, faster than any human and far more accurately, given his natural olfactory endowment."

  Cory paces in front of his whiteboard. "I need to continue building matrices of wines, teasing out which parameters matter to Snoots, which don't. But, if Snoots can find more matches between inexpensive wines and Pillar's Picks, then Snoots is our killer app!"

  "Our killer app? You trained him. It's pretty breathtaking, actually." Dawn stares at him in awe.

  He looks at her more seriously. "This could change things, the way people find wines they like to drink. Even if he just finds close matches between inexpensive wines, it's valuable information to a wine geek. My job will be to let him sniff wines from small unknown wineries, and try to match them with Pillar's Picks from Napa, Sonoma, Oregon, the Central Coast. I could write up my findings on a blog. Hard core wine geeks might actually pay for this kind of information, once they believe it."

  "Will the blog mention that a dog is making these calls?" Rob asks.

  "That's a really good question. I think for now the answer is no. It's a competitive advantage, relative to other wine reviewers, so that makes it a trade secret."

  Rob smiles. "On the Internet no one knows you're a dog."

  Dawn suddenly registers the landscape of wine bottles cluttering the living room. "Cory, are you okay buying all this wine? I mean, without a revenue stream at the moment?"

  He looks down. "Frankly, Dawn, I could really use some investors. The high end stuff is burning through my cash pretty quickly."

  Rob considers this for about a second. "I'm definitely in, Core."

  "So am I," Dawn adds, without hesitation. It's why she brought it up.

  Cory shakes his head, looking frankly at Dawn and Rob. "Wow, are you sure?"

  Dawn flashes a smile at Rob. "Do we look like quitters?"

  "No guts, no glory, man." Rob smiles broadly.

  Cory looks his friends in the eyes. "I won't promise it's without significant risk, but whatever the blog, the business brings in, we're equal partners."

  Rob laughs. "What about Snoots? What does he get?!"

  "Peanut-butter dog biscuits."

  "Have you thought of a name?" Dawn asks.

  Cory pauses to consider this. "What about just Blind Tasting?"

  "It's accurate," she agrees, smiling. "Unless Snootsy can read the labels on wine bottles."

  Rob goes to the kitchen counter, opens three Anchor Steam beers and carries them back, handing one to Dawn and one to Cory. "To blind tasting!"

  "Blind tasting!" they echo, clinking bottles.

  Snoots, panting lightly, watches the three of them with his wolfy smile. The pack seems content and everything is in order according to his high shepherd standards.

  Chapter Sixteen

  Cory continues to build matrices of matches and non-matches for all the wines Snoots is sampling. For each sample set, Cory exhaustively permutes which wine is the control and verifies, importantly, that changing the control within a given set never alters which two wines the dog matches -- if a match occurs.

  In the context of so many tasting trials, Cory succeeds in eliminating the voice command 'find it'; offering Snoots the control has now become a sufficient prompt by itself.

  There are also interesting occasions when a wine, W1, from set S1 matches a control C, but then is not selected when W1 is included in a new set S2, when Snoots chooses a new wine, W2 over W1. This supports the idea that Snoots has a 'threshold of similarity' that must be met in order to match two wines, but that the threshold can be exceeded by an even closer match later discovered. These are rare occurrences, but they offer an extremely valuable clue as to what signals the dog is responding to. Fertile ground for more empirical investigation.

  Another critical finding is that Snoots doesn't always make a match between a control and some wine within a given sample set. This supports Cory's belief that the dog is finding real clues in the scents of the wine, and not just matching random wines in order to receive a treat.

  Finally, a deeply important result is that Cory has never rejected a match that Snoots has made; his own human palate can detect the similarities. But, it's one thing to have the dog rapidly pinpoint these for you, the needles in the haystack, but quite another to go looking for them yourself. Cory knows his own palate would become incoherent with fatigue if he attempted to explore wines at this depth and extent without Snoots' acute canine nose leading the way.

  Cory notes that the situation resembles NP-complete problems in computer science; these difficult problems require non-deterministic polynomial time to compute their solutions. However, if the solutions are found, they can be verified very fast in deterministic polynomial time. With respect to the wine, Cory doesn't have to solve the problems himself, but only check the results that Snoots provides him.

  Cory's Living Room. "So, what do you think?" Cory orients his laptop so that Dawn and Rob can view the screen.

  Dawn's blue eyes scan the home page of Blind Tasting, the blog. She reads aloud, "Left Coast vintages: the good, the great and the unknown." She smiles at Cory.

  "That's the mission statement," he asserts proudly.

  "I like the simplicity. The wine glasses are kind of inviting. Puts me in the mood." Dawn laughs, pointing at the jpeg images of two glasses of wine -- one red, one white -- juxtaposed over a black background. They tilt toward each other in a toast, and the name Blind Tasting appears above them in a contemporary script of eye-catching deep burgundy with golden-white reflections bouncing off several of the letters.

  "In the mood?" Rob repeats, raising an eyebrow at her.

  Dawn’s smile widens, but she can't look directly at Rob without betraying her vulnerabilities.

  "With a few keystrokes we're online to the world," Cory says. "Only trick i
s, we need to get noticed."

  "So how many wine blogs are out there?" Dawn asks.

  "The fraction of web-space that concerns itself with wine is tiny. Pillar's blog is among the most widely read, but most of the others are just Google fodder, they don't get zip for traffic."

  "Which means they don't make coin," Rob concludes.

  "Right."

  "I've looked at a few of them,” Dawn admits. "Endless text describing different wines, one after the other. It makes for god-awful reading compared to juicy sex gossip or political rants. You need to create a buzz. But how?"

  Cory smiles at her. "Be as famous as Leonard Pillar. He's the gold standard, regardless of what the critics say. A winery whose product achieves anything above ninety Pillar points becomes visible on the wine map of the world. Wine geeks will pay money to read the opinions of a guy with that kind of clout."

  "What's Plan B?" Dawn leans her head back against the sofa, trying to conjure up a Plan B.

  "I'm going to include all the similarities that Snoots finds, not just those between Pillar high fliers and unknowns. And, I'm going to display these similarities graphically, plotting their attributes relative to each other. I'll have links side-by-side to the wineries whose wines I'm saying are similar. I'll invite readers to rate my similarities and post those results." Cory smiles. "The only hidden piece of this empirical presentation is how the shared attributes were found in the first place."

  "Do you think people will pay money for this information?"

  "The only way to know is to try, Dawn. I'll put some teasers on the home page accessible for free to establish my, well Snoots', credibility. Like the Grouse-Todd French thing. And I'll be reviewing some wines, but with a twist. If readers want to know what else tastes close to the one being reviewed, they click through via PayPal to a restricted page that shows them the similar wine with the charts and other information."

  "So what do you charge?" Rob asks.

  "Probably something in the iPhone apps range. I'm thinking I'll calculate different prices depending on the spread between the prices of the wines I'm comparing. For wines that are similar and both low-end, I'd charge the ninety-nine cent rate, for wines that are similar and both expensive, I'll charge a bit more, and for wines where one is low-priced and the other, high-priced, I'll charge the most. It'll serve as a code to the readers to know the kind of information they'll be accessing."

 

‹ Prev