Finding wines that you will probably like: a step in the right direction

A company has created software which aggregates your and other wine tasters' scores on a wine, and uses the data to predict whether you will like certain other wines.

That is a step in the right direction, but a better predictive product would merely use our personal preferences for the following spectrums:

Fruit Ripeness: from banana/green/peppery through red fruits, through purple fruits, through black fruits, to raisins and tar.

Body: from thin to rich/fat

Residual Sugar: from bone dry to super-sweet

Acid (quantity of acid): from flabby to sharp

pH (strength of acid): from weak to strong

Complexity: from simple to layered

Aromatics: from none to faint to lush olfaction

Finish: from none to long

Tannins: from young and sharp to aged and subtle

These would be adjusted differenly for white and red wines, of course. And some of those spectra are matters of personal preference, whereas others are more clearly "good" or "not as good," but you usually can't find a long finish in a super-cheap wine. At any rate, once we knew our own preferences on those spectrums, then if each wine label were labeled to show those spectrums with that wine's marker in each category, we would have a good chance of knowing if we would like that wine. We will always know the price, of course, so we can compare price to the various qualitative data points, and voila! we are out of the wine woods and we are an educated wine consumer (as to that wine).

Compare this pipe dream to today's wine label, which often tells us almost nothing about the above criteria. The winemaker knows all that info; why not share it with us?

You can read the article here.