Get a glass of wine ?

With wine, with poetry,
or with virtue, as you chose.
But get drunk.

: Charles Baudelaire, Paris Spleen

There is an abundance of opinions that will tell you the right wine to drink to match the season, the food and the colour of your shirt.

But how much do people really know about wine?

In 1976 a British wine merchant organised a wine tasting to prove once and for all the superiority of French wine. It is now popularly known as the “Judgement of Paris”. The expert judging panel was pro-France too with 9 experts from France and 1 each from USA and UK. To their collective horror, in a blind taste test the experts ranked a Californian wine as the best in both the red and white wine category.

In 2001 French acedemic Frederic Brochet offered the same wine in two different labels – once as a cheap table variety and once as an expensive grand cru. His subjects described the grand cru in flowery terms and the table variety as weak and flat. It was the same wine.

Cut to 2011 where a British psychologist, Richard Wiseman, ran a simple experiment. He bought a lot of wine priced between $5 to $50 and then asked 600 or so participants of a conference to say which one was more expensive. The experiment was run double blind – neither the person offering or the one drinking it knew which wine they were drinking. People could pick the expensive wine only 53% of time. Coin flip, anyone.

Weight opinion and expertise by domain specific probabilities. Science and medicine are high probability domains for known knowns. A doctor will be right 99% of the time for commonly found symptoms. Outside of a few exceptions, most physics and math problems do have closed form solutions. The experts really know the right answer. However matters of taste like art and drinks are different. Expert in such domains are apostles of current and past fashions than arbiters of any universal truth.

So what about investment experts and stock pickers? I would give them slightly more credence than art and wine experts, but nowhere close to the sciences. It is not because they are not smart. It is just the highly probabilistic nature of the domain they are operating in where you can make out like bandits on a bad thesis and loose your shirt on a good one. Keep that in mind.

As an aside, I hope one day there will be a similar Judgement of Palo Alto, where to their surprise they will find software and products made in Bangalore are better designed than those in California. And we will all drink to that!