Do the En Primeur scores predict well the future quality of wines?

In Bordeaux, En Primeur wine scores are traditionally released from April to June. The experts taste and assess the wine in barrel. Their scores represent the wine potential after ageing in barrels, once it is bottled. It is a prediction of the wine futur quality.

Once the wine is in bottle (usually 2 years later), the critics re-rate wines and continue year after year as wine ages in bottles.

As the experts’ En Primeur ratings have an influence on wine market and wine prices, it seems interesting to observe this score evolution from En Primeur period to wine aged in bottles. This may show how good are the En Primeur scores in assessing the potential of each wine.

The study

This study is done using the Global Wine Score database which gathers critics ratings (from 20 critics). Only Bordeaux wines are concerned from the vintages 2005 to 2010. Studying period before 2005 would not allow to have a homogeneous analysis, because of a too small amount of critics data in our database. Concerning vintages after 2010, wines haven’t aged enough yet, so it is difficult to observe a true evolution.

The charts present the deviations calculated between the re-notation and the En Primeur score for all wines from a given vintage and having re-notations for the year concerned. When this deviation is positive, it means that the average re-notation is higher than the En Primeur score. When the deviation is negative, the En Primeur rating is higher than the re-notation. The linear regressions complete the charts by giving a trend for each vintage.

The analysis

The data of this study has been statistically processed to exclude outliers.

Different cases exist concerning the vintages shown on the chart. It seems interesting to put in perspective this study with our blog article which compares the different vintages (see our blog here).

On the charts below, blue dots correspond to deviations between re-notations and En Primeur scores. The red line corresponds to the linear regression for these re-notations with its equation in the down right corner.

Vintage 2005

Concerning the vintage 2005, from 2007 to 2017, the linear regression is almost steady with a very slight decrease. It means that re-notations by critics are globally very close to the En Primeur scores.

Vintage 2006

2006 is an classical vintage in Bordeaux. The decrease of the linear regression is clearer. According to the critics, wines may lose around 1 pt every ten years, re-notations being lower than En Primeur scores.

Vintage 2007

2007 is admitted as a “bad” vintage by wine critics in Bordeaux. In this case, the linear regression is steady, the re-notations being very close to the En Primeur scores.

Vintage 2008

2008 is a classical vintage in Bordeaux and, like vintage 2006, there is a slight decrease of the linear regression with years, equivalent to 1 pt every 10 years.

Vintage 2009

2009 is a “very good” vintage in Bordeaux and the linear regression is steady.

Vintage 2010

2010 is also considered as a very good vintage, but the decrease is very clear. The En Primeur scores on average may lose 2 pts every 10 years!

The table below tries to sum-up all these results.

The table shows the reputations of the different vintages from 2005 to 2010 associated to the average Global Wine Score for each of them and the trends resulting from this study.

The first interesting finding is that, from 2005 to 2010, two situations appear. Vintages 2005, 2007 and 2009 are more or less steady while 2006, 2008 and 2010 are decreasing.

The vintages 2005 and 2009 are generally admitted as very good vintages in Bordeaux by wine critics (confirmed by the respective GWS averages 91.8 and 92.51). 2007 has a lower reputation (average GWS at 89.34) and has also a steady trend.

On the other hand, vintages 2006 and 2008 are considered as common vintages (respectively average GWS 90.21 and 90.69) and have decreasing trends of 1 pt every 10 years. The strongest decreasing trend is vintage 2010 where wines may lose 2 pts in ten years while this vintage has a very good reputation (average GWS at 92.5), if following the prediction trend.

Conclusion

As Bordeaux wine trade is based on negociants’ system where critics have a great power, an analysis of the accuracy of the En Primeur ratings is crucial. This study gives a new point of view on that system.

Globally, re-notations are very close to En Primeur ratings and this study tends to show that finding. That means that wine critics have a good prediction of the future quality of the vintages.

The period studied from 2005 to 2010 include three years admitted as great vintages: 2005, 2009 and 2010. Vintages 2005 and 2009 are steady while 2010 has an important decrease. Nevertheless, this great vintage seems now to be in par with 2009.

This study is not intended to give a perfect and universal vision about these vintages from 2005 to 2010. But it shows the Global Wine Score potential for data study in wine. As databases will grow bigger and bigger in the next decades, it will be great to do the same study when periods available will be 20 or 30 years long. As weird as it may look like, mathematics offer a new point of view in assessing wines.

Time will tell…