Whom among us has never, at least once, had the unpleasant experience of opening a promising bottle of wine, only to be severely disappointed – even though the tasting notes of our preferred wine expert were altogether enthusiastic, even dithyrambic – and then wondered what to do with the other eleven bottles in the cellar? This is all the more vexing now that fine wine costs double, triple, or even quadruple what it used to! Or, maybe you have experienced the frustration of discovering, while reading a well-known wine guide, that you missed a golden opportunity to buy an excellent wine that is now almost impossible to find?
Fortunately, the art of wine tasting is not an exact science. The experts are, after all, only human. Each one has their particular strengths and weaknesses, and their appreciations also vary according to their physical and mental state at the time of tasting. Furthermore, the conditions under which a tasting takes place (early in the morning or late night, blind or not blind, noisy or silent, in a well-lit or dark setting, whether the samples represent wine that has already been bottled or not), as well as the influence of other participants, the way the wines are presented, the quality of the glasses, etc. can significantly influence a taster's impressions, although he may not realize it. For obvious reasons (a lack of time and money), it is practically impossible for the average Bordeaux lover to taste several hundred wines every year to form his own opinion. Therefore, many consumers rely on the informed criticism of professional tasters, wine journalists, sommeliers, wine retailers, or other enthusiastic wine lovers in the wine press or on Internet forums. However, this entails the sorts of risks mentioned above !
After reading various wine guides and magazines for years, I noticed how frequently – and how greatly – scores can differ from one taster to another for the very same wine. I found this all rather frustrating, so I set up a database in 1997 making it possible to compare the tasting notes of a number of experts, and identify developments (in terms of quality) at the various châteaux. Wishing to create a simple and unbiased way of providing an accurate representation of wines and vintages, I thus created this site. It is based on calculating an average score for each wine, taking into account both similarities and differences of opinion, and measuring the standard deviation.
The first problem was comparing various ways of rating wines (there is no ISO standard for this!). There are, roughly, two distinct approaches to scoring: the "relative method", and the "absolute method". The relative method involves comparing a sample to a reference wine, generally the best of a given group. The absolute method is based on comparing wines to a hypothetical perfect benchmark, whatever the wine or vintage. Most critics use the absolute method. As regards rating scales, there are a number of different systems (a number out of five stars, a score out of 20, a score out 100, etc.). The most famous American tasters – Robert Parker, Stephen Tanzer, and James Suckling – give marks out of 100. This American approach is extremely precise as it allows for subtle differentiations. The European approach, a score out of 20, is generally more flexible, the aim being to give an overall impression of the wine rather than attempt any sort of scientific assessment. For instance, is a wine rated 89.5 really that much better than another that scored only 88? It’s all a matter of taste…
For the sake of convenience and efficiency, we have adopted a 100 point rating system for our database. The other systems are therefore converted to scores out of 100 in order to be compared. Please note that for American tasters, a score of 50/100 is, in reality, the extreme lower limit in terms of quality.
The scores we attribute are an average of all the marks given by expert tasters. However, we had to be careful to avoid certain pitfalls in interpreting these scores. Take, for example, two wines, both of which score an average of 88/100.
| Score 1 | Score 2 | Score 3 | Average | |
|---|---|---|---|---|
| Wine A | 86/100 | 90/100 | 88/100 | 88/100 |
| Wine B | 81/100 | 97/100 | 86/100 | 88/100 |
In the case of Wine A, we are dealing with closely-bunched scores, reflecting tasters who shared a very similar opinion.
On the other hand, although Wine B had the same average score as Wine A, it received three very different evaluations. The first taster did not really like this wine at all, the second taster loved it, and the third is fairly neutral. Therefore, a score of 88/100 does not really give an accurate picture of the quality of this wine.
This is why we incorporated an additional factor into our averages: “standard deviation”. This statistical tool is often used in finance (as well as in certain scientific fields and the social sciences) as a precautionary measure. Standard deviation measures the spread of values as compared to the average. With regard to our analysis, the higher the standard deviation, the bigger the difference between scores.
In the preceding example, the standard deviation for Wine A is 2.0, i.e. a result relatively close to zero (zero = no difference between the three scores). For Wine B, the standard deviation is 8.2, since the dispersion of scores is much more widespread than the average score of 88. This does not necessarily mean that Wine B is not as good as Wine A, but simply that, due to certain circumstances (such as a taster not appreciating the wine's style), the reasons behind such a disparity bear looking into.
As of today (*), our database includes a mass of useful statistics – more than 25,000 scores from wine experts! Based on a multitude of objective and indisputable parameters, we have compiled these conclusions on StatWine.com to help you buy your wine intelligently. Our database indicates:
Since statistics alone cannot replace the tasting experience, we also include tasting notes for each wine. These will be regularly updated.
Generally, wine is intended to be enjoyed during a meal. Therefore, we also suggest possible food/wine matches for each wine.
In addition to detailed technical information for each château (such as breakdown of grape varieties, area under vine, viticultural methods, vine density, average yields, number of bottles produced, barrel aging, etc.), we offer an objective, concise analysis to help you buy confidently. The "When to Drink" section of this site is also highly useful to find out when the treasures in your cellar are ready to drink.
Furthermore, easy-to-understand graphs show at a glance if a château is better, or not quite as good, as other wines in the same appellation. You will also be able to visualize the general trend in scores over the past ten years and see the progress made by certain estates.
This website is the result of many years of patient work, and we are still full of ideas for developing StatWine, a site which provides one of a kind service to consumers and wine lovers.
The StatWine.com team welcome you to our site, which we are convinced will rapidly become a reference for Bordeaux wines.
On behalf of StatWine.com,
Alain Bringolf
(*) January 2008