The other day, I was working with my D&D group to perform a regression analysis on Wine Spectator scores. In order to select the variables (i.e. price, region, specific tasting notes) that were most relevant/important in determining the final rating, I had to understand the score structure. I thought I would share some of these learnings - all of which were new to me.
There are many different types of wine scores available - Wine Spectator, Wine Advocate, and Wine Enthusiast. Wine Spectator, whose data I used, notoriously gives out the lowest ratings.
Diving into historical Wine Spectator scores from 2009-2013, Wine Spectator gave out an average score of 88. Few wines scored below 81 and very few wine scored above 94. That means that the “average” 88-point wine is considered “good to very good,” and few wines are “flawed and taste average” (a score below 80) and “benchmark examples or classic (a score above 94). One critic writes that wines scoring below 70 simply wouldn’t be reviewed at all.
When looking at the specific score rubric for Wine Spectator, almost all wines immediately get 50 out of 100 points for simply being wine. From there, a variety of factors, listed in order of most important to least important, affect the score in the following ways; “flavor and finish” taste of wine adds up to 20 points, "aroma and bouquet” adds up to 15 points, "overall quality" adds up to 10 points, and "color and appearance" adds up to 5 points.
In addition, many wine industry experts write that the scores are specific to wines of a certain varietal. For instance, a 90-point score for a Cabernet Sauvignon wine will not necessarily correlate to the same quality experience for a 90-point Riesling wine.
Most interestingly, however, was the fact that wineries specifically create wines based on the taste profiles of famous tasters. “Parkerization” is a phenomenon where “wines [are created] to match the preferences of famed wine critic Robert Parker” (Inc). And, when we compared our data results to the reviews of specific tasters, our group noticed trend correlations to specific reviewers, as well.
And last but not least, some of the key words that are important in terms of determining scores were “complex,” “balanced,” and “structured.” The appearance of the word "fruit" is so far, insignificant, in terms of the overall score.
Sources Used:
No comments:
Post a Comment