Wine ratings 2 of 2

Wine ratings engender security.

Plunk down $85, you want to ensure your plunk didn’t buy plonk.

But, do wine ratings—or their fellow travelers, medals at wine competitions—really mean anything?

When Americans were in their neophyte wine stage, way back in the Neolithic 1980s, no one wanted to be a wine rube. If Robert Parker or Wine Spectator blessed a bottle with their imprimatur, you could safely spend and pour and impress your friends.

Two things happened. First, global explosion of good wine means it is much harder to buy bad. Second, former wine fledglings now have consumed many, many cases of wine and can fly. We know what floats our wine boats, and we don’t need a fat former lawyer or shiny slick magazine to tell us what we like.

Give wine scorers credit. They consistently assert every palate is unique and good wine is what you enjoy, you do not have to listen to me. We embrace the premise.

Study after study reveals capriciousness of scores and competition medals. A retired college professor and small-time winemaker, Robert Hodgson, joined the California State Fair’s wine competition advisory board. He convinced board to run a scientific study of tastings. They agreed, so long as he kept findings confidential. Do not pull back the curtain on the Wine Wizard of Oz.

You can easily guess what Hodgson did: poured from same bottle at test after test. Scores bounced around from 85 to 95. Very best judges rated the exact same wine plus or minus two points, other scores varied by as much as 10 points. The most common difference was four points. For the exact same wine. Oops.

Hodgson wrote up his studies each year. Each year the wine competition board refused to allow him to publish. Finally, they relented. Finding: chance of winning a medal in competitions “mirrors what might be expected should a gold medal be awarded by chance alone.”

In deference to such research, no wine recommendations this week. Drink what you like. If you enjoy it, it is good. That even goes for white zinfandel. For MD 20/20 or Thunderbird, not so much. We must draw the line somewhere.