If not scores or words to describe wine quality, then what?

The point that I have been making in a number of my recent blog posts is that wine-quality scores give the illusion of being mathematical without having any useful mathematical properties. This is not quite fraudulent, but it is unfortunate — the apparent precision of the numbers gives an illusion of accuracy.

This means that, in practice, quality scores do nothing more than express personal preferences — they allow the critic to put wines in some sort of rank order. There is nothing wrong with wanting to do that, of course (see Wine tastings: should we assess wines by quality points or rank order of preference?). Indeed, if you can find someone with the same tastes as yourself, knowing about their preferences can be very helpful for buying wines.

However, this does raise the obvious question as to what we might use, instead, to express a rank order of preference. This question has been raised by a number of people who have expressed their conclusions publicly, as well as by many who have not. This blog post gathers together some of the public ones. The collection is neither comprehensive nor exhaustive, but merely introduces you to some of the suggestions that I have found interesting.

Using letters not numbers

This approach is what teachers use, whether they are at a high school, a university, or a community college. So, you are all familiar with it. Each student's academic score (expressed as a %) for each semester or year-end is converted to a simpler grading system, which then expresses the student's achievement.

Sometimes, the grade ranking is something like this (from best to worst): A, B, C, D, E and F, where D conveniently means "Deficient" and F means "Fail". Other times it might be like this: A+, A, A-, B+, B, B-, C. The one I had at university in Australia used: High Distinction, Distinction, Credit, Pass, Terminating Pass, and Fail. Note that in this case D means "Distinction", not "Deficient"! There is nothing universal about educational grading systems.

This general approach is thought to work better than the original percentage scores because there is an explicit meaning to each of the grades — getting an A means something quite different to getting a B. This forces the teacher to be very careful about which students get which grades. As noted by David Schildknecht:

Scores of wine critics aren't bound to intuitively meaningful letter grades as are those of educators — although I grant that the association of a 100-point scale with educational grading is so strong that even in wine criticism, despite incommensurability, many consumers will think of e.g. "90" as having some intuitive sense.

This grading system can work provided that the institution involved has a convention for assigning a particular numerical range to a given letter grade, and that the grades objectively correspond to a specified level of academic achievement,

Moody credit rating system

This idea was suggested by wine merchant David Farmer (John Moody and brave or foolish judging). David was unhappy with current wine judging systems, particularly the well-known fact that a wine can win a gold medal at one show and nothing at all at the next show. He was therefore looking for a more complex judging system, which he calls Three Dimensional:

After a bit of searching I stumbled across the inventiveness of John Moody who in 1900 developed a rating system for financial paper. I found his ideas had a three dimensional character to grading which offered, if not solutions for wine judging, a display of technique which could perhaps be adapted ... As I studied the Moody system I found it offered many parallels to wine judging.

Moody's rating system is one of several that are used to express credit ratings in the financial world (the others being from Standard & Poor, and the Fitch Group). It can be applied to bonds and to countries, for example, to indicate their credit-worthiness. The initial idea was to provide to investors impartial information on credit-worthiness of financial entities, which the credit ratings agencies were being paid to provide.

We do not need to concern ourselves with the methodology used for the economic evaluation, but only with the way the outcome is expressed. The Moody system uses a linear ranking with 21 levels (from best to worst):

A A A A A A A B B B B B B B B B C C C C C

a a a a a a a a a a a a a a

a a a a a a a

1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3

Clearly, the system is more complex than a simple score of 0 to 20, such as might be used for wines. In the Moody system, it matters whether the entity being rated is in group A, B or C, and within the groups it matters how many "a"s it gets. For example, being downgraded from Aaa to Aa1 has a specific meaning within the system (see Wikipedia). This is in contrast to a wine-scoring scheme, where adjacent scores have little or no objective meaning.

This general approach to expressing a rank order has a lot of similarity to that of the educational system, as described above, being at heart a more elaborate version of it.

Wine tally

James Ho uses the name WineTally on Facebook, Instagram and CellarTracker. He has suggested what he describes as "a Simple, Sensible Scoring System for wines to enable every consumer to articulate one's evaluation and experience." This boils down to providing the reader with the individual sub-scores as well as the final score.

There are 4 wine-character-dimensions (Accuracy, Balance, Complexity, Depth), with 3x4x4x3 = 144 possible outcomes. Technically, the wines are thus placed into one of 144 points in multi-dimensional space (ie. there are only 144 possible different wines). These 144 outcomes are then reduced to one of 11 possible scores (scale 0-10).

Most of Ho's wines get a score of 7, 8, or 9, occasionally 6, rarely 5 or 10 (see the examples below). There are 18 ways to get a score of 7 (ie. 18 distinct wines could get a score of 7), ten ways to get 8, and four ways to get 9. So, this certainly keeps things simple, in terms of expressing a rank order of preferences.

We might see this as a simplification of the previous two systems, but at the same time providing more detail.

Music rating

Bob Henry has suggested using a form of reviewing that already exists for assessing music. He has noted:

Audiophiles who read British review magazines have been introduced to a twin scale: one judging the sonic fidelity of the recording (against the reference standard of live acoustic music performed, without electronic amplification, in a concert hall), and artistic interpretation of the musical piece.

As an extreme example: over 100-year-old Edison phonograph cylinder recordings of operatic tenor Enrico Caruso have low sonic fidelity (by contemporary standards) to the sound of the human voice and accompanying acoustic instruments ... yet could represent the finest artistic interpretation in recorded history. The recording gets two scores: an “F” letter grade for [poor] fidelity and an “A” letter grade for [great] artistry.

Similarly, a wine could be less than true-to-type or even technically flawed (50 points) ... and yet be hedonistically sublime (100 points).

Thus, this expands the wine score to two separate components, one technical and one artistic, rather than the usual approach of confounding the two. Each wine gets two scores, which may be numerical or not.

Further thoughts on music

The connection between wine and music has been taken much further by other people. For example, Morten Scholer has noted:

Some winegrowers, in particular in Italy, have loudspeakers in their vineyards and in their cellars playing classical and baroque music. They have experienced that the sound waves have a positive influence on the quality of the grapes and wine.

He lists a few specific examples; and David Schildknecht adds another (2012. The World of Fine Wine, issue 36, p. 32). This idea has been formally examined by Charles Spence and Qian Wang (2015. Wine and music (I): on the crossmodal matching of wine and music. Flavour 4: 3).

Of more relevance to the discussion here, Michel Bettane has suggested (The music of wine):

For me, the individuality of a noble wine is analogous to music-tone, voice, song. I think of wine as a musical score, composed of geological, climatic, agronomic, and enological notes that must be sight-read, understood, and interpreted. The taste is the result of that interpretation.

The Australian wine critic Mark Shield took things even further than this. He used an alternative approach to that of numerically scoring wines, by comparison to music itself — that is, the music that the wine most resembles, or the music it makes you wish you were listening to. This is analogous more to a wine note than to a wine score.

Philip White has noted:

Mark Shield, the dreadfully missed Melbourne wine writer, made no bones about how the music of Thelonius Monk influenced his palate as he reviewed wine. He said he couldn’t taste properly without it, and often referred to different compositions of his favourite jazz composer and player relative to specific wines.

As but one specific example of Shield's approach (using classical music):

If they ever bottled Beethoven's 5th Symphony it would be the 1982 Cape Mentelle Cabernet sauvignon.

Shield was a widely admired writer in a self-confessed Australian larrikin style. It is a great pity that his wine columns (Rough Marc in Wine and Spirit, Noble Rot in the Sunday Age, as well as others) are not freely available on the web, for all to read. His missives from the Rat Shack are among my fondest memories of my first forays into the wine literature.