Global Wine Medal Rating (GWMR). An introduction

Methodologies to assess wine: A labyrinth of diversity

We are confident it is no secret that there is a multitude of wine ratings, whose fundamental purpose is to assess the wine quality in a clear and most importantly, comparative manner. Which is, using a certain method of wine assessment (organoleptic, as a rule), to give the wine a numerical score in points.

Such an assessment gives the final user of this rating (usually an ordinary customer) clear information about which wine, according to experts, is better and thus, grants a certain degree of awareness when making a decision to buy.

We insist on avoiding to raise the issue of which of the dozens of respected and proven methodologies is more correct and which of the assessment scales reflects the quality of the tasted wine in a more adequate manner. We would simply like to underline that such an assessment would be quite subjective at all times. This is why ratings have a quite specific authorship, while for everyday use the edition, for which the wine critic, who assessed the wine, works is usually indicated.

Alternative ratings: Ariadne’s threads

By far not every wine expert remembers the differences between various assessment methodologies in all nuances. It is even more difficult to compare the ratings of wines if one of them was rated on a 100-point scale, the second on a 20-point scale, and the third on a 12-point scale. It is for this reason that the tendency to create “aggregated ratings” has become more and more clearly manifested, seeking to reduce many close to each other ratings to a single denominator, or to take into account other factors, not directly related to the sensory analysis of wine, in the final assessment.

We shall give just a few examples, not trying to weigh their pros and cons, but only illustrating the versatility and sometimes creativity of approaches:

• Vivino (5 ☆) — An aggregated rating of non-professional ratings of the users of the application with the same name.
• Global Wine Score (100 points) — One of many aggregated professional ratings.
• Reverse Snob Rating (20 points) — A method that takes into account the derivative of the quality and price of wine.
• Predicting wine quality by its composition (using machine learning technology) (100 points) — A method for assessing wine quality based on 11 physical-chemical parameters.
• AI prediction Rating (100 points) — A self-learning algorithm for calculating wine quality score based on the length of the comments accompanying the assessment (the longer and unique the comment, the better the wine).

We should also mention our fellow nationals from WineStatistics.com, developing the WS.Regressor, a machine learning algorithm, that allows predicting professional wine ratings. We are looking forward to the results.

Accuracy issues: A wrong turn?

In general, while supporting the trend of aggregated ratings, it is still worth emphasizing that all of the above methods either operate with indirect data or take the same subjective assessments as a basis.

In addition, any, even professional assessments by the most famous wine critics, have three fundamental issues:

• In most cases, the taster KNOWS what kind of wine is being assessed, which cannot exclude the influence on the final assessment of the credibility of the brand, region, etc.
• In most cases, the taster assesses the wine individually, and not in comparison with a representative list of potential competitors.
• In most cases, the assessment is performed by one specific expert, and not by a representative panel of tasters.

Is there any other source of information: Shall we take off?

It would seem that there is a very large-scale source of reliable information on collegial and, more importantly, blind assessment of wines within the framework of representative samples of the best wines from many representative winemakers from different continents, countries and regions — awards of significant international or national professional competitions.

Moreover, winemakers, distributors and wine merchants spend significant marketing budgets to participate in these competitions, highly appreciating potential results and clearly valuing received awards.

Nevertheless, collecting, structuring and analyzing information about wine awards at professional competitions is far from simple. That is exactly why at the moment there is not a single aggregated rating based on the medals of a wine. When working on such a rating, we would have to take into account the following:

• There is an enormous amount of information — more than a hundred thousand wines receive awards annually at more than 400 competitions.
• The scale and significance of the competitions are very different. It is necessary to identify and calculate clear comparative parameters of competitions.
• The titles of the award values vary from competition to competition, which is an additional task to systematize and grade them. Assessment ranges for medals of similar merit differ from competition to competition.
• The very fact of receiving a certain medal does not indicate the exact assessment of the wine, but only the range within which this assessment fell.

Will it work out: Icarus or Daedalus?

Gustos.life project team decided to develop another alternative rating option based on its own long-term experience in providing professional wine competitions by means of the software: https://winetaste.gustos.life/index.html.

The idea is to create a wine rating based on a mathematical model that reflects the success of the participation of a particular wine vintage in international professional wine competitions.

The Moldovan wine market, for which we have information about all the medals received by Moldovan wines at all professional competitions in the world over the past three years, was chosen as a pilot for testing the model.

For 3 months, the project team developed and tested a mathematical model, which we called the “Global Wine Medal Rating” or the GWMR Index.

While working on the rating, we created a system for an automated collection of information about the medals of all wines from all countries at all competitions in the world, as well as a self-learning algorithm for structuring and analyzing information about medals and wine competitions, so that the rating becomes truly Global.

In the next article, we’ll describe in detail the mathematical model of the GWMR Index, including a description of our approach to creating a global rating of professional wine competitions.

And later, in a separate article, we shall describe how the results can be used in a practical way:

1. By wine lovers, especially in terms of the rating undervalued wines.

2. By wine companies when choosing the best strategy for participating in wine competitions.

3. By wine sellers, especially online shops, in terms of presenting both the wine rating, as well as comprehensive information on all medals of a specific wine (vintage).

Authors:

• Andrei Radul (Linkedin profile)
• Veaceslav Cunev (Linkedin profile)

About the project Gustos.life:

The Gustos.life project specializes in providing information services for the wine industry within the created integrated ecosystem.

Main Services:
1. Service for holding professional wine competitions — https://winetaste.gustos.life/

2. Digital wine storage — https://dws.gustos.life/