Last night, a dozen or so friends and I discovered one of the best-value wines in America.
I’ll tell you what it is in a minute. But first, it’s worth explaining how we came to that conclusion: Michelle and I held a Pinot Contest at our apartment. The structure was pretty much the same as the wine contest we held last year: each contestant brought two identical bottles of wine. One was tasted blind, while the other was kept for the prize pool. This year, every bottle had to be either a Pinot Noir or a Burgundy.
Everybody scored every wine out of 20, and we added up the results, which you can download as an Excel spreadsheet with a wealth of information in it.
With 14 people tasting 12 wines, the maximum score was 280. In the end, the scores ranged from 115 to 212. And it’s worth emphasizing that the overall quality was extremely high: much higher, actually, than it was last year, when people could bring any red wine they liked. If you make a bit of an effort with a Pinot, it seems, you’re likely to be well rewarded.
But the other thing which really sprang out from the tasting was that there was absolutely no correlation whatsoever between price and quality. Here’s the chart:
If you plug these numbers into a correlation calculator, you actually get a negative correlation of -0.2: the higher the price, the worse the wine, on average.
As you can see from the chart, the real standout wine was Wine F, which got the best score of the evening (212 points) while costing just $13 per bottle. It’s not a typical Pinot: it was full-bodied, and fruity, and utterly delicious. “A Burgundy probably,” wrote Jay on his scoring/tasting sheet, adding “expensive” and scoring the wine 20/20. “Simon?” wrote Seth, knowing that Simon had brought a very expensive Burgundy, and also giving the wine the full 20 points. “Yummy,” wrote Gaby, giving it her top score of 17. When she poured too much of the wine into her glass by mistake, she asked for a second glass: it was simply too good to pour away, but she did need to carry on tasting.
Wine F turned out, to everybody’s surprise, to be the 2005 Heron Pinot Noir, made in California from French grapes by Laely Heron. It cost just $13 a bottle, or $11 if you buy by the case — which I assure you I am going to do. Meanwhile, Simon’s expensive Burgundy (Wine D) managed to get a total score of just 146: only three wines scored lower.
This contest was emphatically not a triumph of cheap New World wine over expensive Burgundies. In the bang-for-the-buck ratings, the top wine in terms of points per dollar was mine (Wine I), a 2004 Burgundy which got 180 points and cost just $9.99 at Warehouse Wines on Broadway. Simon gave it 20/20, while giving his own $52 Burgundy just 12/20. The Heron was in second place in the bang-for-the-buck ranking (it was made with French grapes, of course), while in third place was Seth’s 2005 Burgundy (Wine K) which also cost $13, and which received a score of 154 points.
At the other end of the scale, Wine J was a $50 El Molino Pinot from Napa, which scored a fair-to-middling 167 points, making it the second-worst value after Simon’s Burgundy. As you might expect, in general the cheaper wines did better in the bang-for-the-buck stakes, but it’s worth noting that Michelle’s Pinot (Wine G) — which came from Germany, of all places, scored 193 points and cost $26 — ranked higher than Gaby’s $18 Pinot Nero from Italy (Wine A), which garnered a mere 115 points despite the fact that I, personally, loved it.
In general, the Burgundies were clustered right in there with the non-French Pinots. While Burgundy has a reputation for being very expensive and rather unreliable, in fact its wines seem to be consistent with Pinots globally. And the second-place wine was a Burgundy: Wine B was a 2006 Alain & Julien Guillot Clos des Vignes du Maynes, brought by Rory but chosen by Jay. Since the winning wine was brought by Savannah, Jay’s wife, that meant that Jay and his party ended going home with all the bounty — although they were generous enough to let me keep the two winning bottles.
Thanks to everybody for coming and making the contest so fun and successful. And remember: sometimes the cheapest wines are also among the best.
Update: Josh Reich does some seriously high-end statistical analysis on the scores, and concludes that “price is not a significant predictor of wine score”.
Interesting to see that the WOTN also received the lowest single score of all the wines tasted with the 1 rating from Maggie (Although there were a couple of other votes of only 1 for two other wines.)
However, you may have noted there still is some correlation between price and quality – not for the group as a whole, but for Andrea. Andrea had both the highest r-squared for both price (r-sqrd 0.396129056) and for bang for the buck (r-sqrd 0.502169978)
Great post; the contest looks like a lot of fun. A couple questions, though:
1) Any comment on sample size? Is the data significant?
2) What is range of tastes for the judges? Does everyone tend to have the same experience with wine or like the same kinds generally? Hard to measure this, I realize, but it would be interesting to hear what you think limits of your conclusions are.
Billy Ray — Thanks for asking. My feeling is that wine tasting has been comandeered, to a large extent, by professional tasters who may or may not have different taste (either physiologically or psychologically) to most of the rest of us — see the whole debate about “super tasters” which I shan’t go into here.
So this is a real-world test, with real-world individuals who are most definitively NOT wine-tasting professionals — although there were definitely a few people on the edges of the wine industry at the tasting. It turns out their scores weren’t significantly different from anybody else’s.
As for sample size, I think 14 is significant, you might not, that’s fine. But I certainly think that 14 is more significant than 1 — which is what most wine reviews give you.
on the sample size point: my friend is a barrister and says that it’s remarkable how often juries reach the ‘correct’ conclusion. similarly with focus groups (which tend to consist 0f 6-8 people). the point is not that the relatively small number of people is ‘statistically significant’ in the usual sense of that term, but that certain *types* of judgements are trustworthy even if the number of people is small.
This definitely does save you alot of money. This is such a great idea.
Thanks brother was very nice of page.