notes from the winemaker

That is the question. I have a number of decanters around the house. I use them when I have a wine that needs to be removed from a sediment in the bottle. But I never decant a wine hours before serving it to let it “breathe” and I don’t own one of those thingamajiggys that aerates the wine as you pour through it. One, I don’t buy wines that some think need such treatment to taste “better.” Two, I value the more gradual change in a bottle of wine as it opens up in the glass over the couple of hours it takes us to drink it.

For me the second point calls into question the validity of the premise implied in the first–that aeration actually improves wine. It may change them, but does it really improve them?

This morning Joe Roberts (1 Wine Dude) tweeted about this article in Bloomberg Businessweek where Nathan Myhrvold discusses “How to Decant Wine with a Blender.” Joe’s comment about hyperdecanting: “I know a lot of wine geeks that actually do this.”

My first thought was “really? why?” Is this just another example of our “more is gooder” cultural imperative? Or is it wine geeks feeling the need to insert themselves in to the winemaking process?–”hey look I made it better!” My second thought was “whatever, you bought it, it’s your wine–you can mix your ’82 Margaux with Tab and serve it over ice if that’s what floats your boat. It’s not like you’re shooting paintballs at the ceiling of the Sistine Chapel.”

But careful reading of Myhrvold’s article reveals an interesting nuance. Early on he states that hyperdecanting in a blender “…almost invariably improves red wines—particularly younger ones…” This is a subjective evaluation, and I believe if you put ten people in a room to taste a blended ’82 Margaux next to a bottle decanted just to remove sediment, you’d get some lively argument about which wine was “better.” (Especially if I was there.)

Then comes something interesting: Myhrvold talks about using a triangle test to remove bias. With at least 10 judges, and multiple presentations. It’s not surprising that the former CTO of Microsoft understands the value and utility of the triangle test as a research tool. It would be astonishing, though, if consumers and professional wine evaluators actually started to use the method to eliminate bias in tasting.

From my perspective the best part of the article is when Myhrvold walks back his assertion that hyperdecanting “improves” the wine, when he says that “…hyperdecanting does clearly change the flavor of the wine.” Note: changes, not improves. He follows up with “To determine with scientific rigor whether your tasters prefer the hyperdecanted wine requires a more complex trial called a ‘paired preference’ test, or ‘square’ test.”

Absolutely. And I can tell you from experience that the results of this sort of evaluation are often something like “60 percent of the tasters prefer wine X at the 90% confidence level.” In other words, on average 6 out of 10 tasters (not necessarily the same group of six in each trial) prefer wine X about 9 out of 10 times.

Does this mean wine X is “better”? No. It still means that a group of tasters express a subjective preference for wine X. If 10 out of 10 tasters prefer wine X at the 99% confidence level, perhaps then we would be justified in concluding that wine X really is better.