Historic Score Cards
Score Card Change Log
30/06/23 - Created individual critics charts to allow end users to pick the critic they believe in most rather than looking at overall averages
Liv-Ex historic observations around En Primeur
How do I interpret this?
A z-score is, in effect, the score for the wine relative to a particular critic's average for that specific vintage. A higher z-score is indicative of a higher score for the wine. The z-scores are averaged across all the critics that score a wine, so you may want to look at individual ones for your own palate/preference. Anything that goes above the line of best fit represents better value for money, and anything below represents poorer value for money. Note it is very important to understand the context around these wines - for example, right bank wines are typically produced in lower volume so may be 'more expensive' per point.
Please take all data with a pinch of salt - I've not tried to apply fully rigorous techniques to this, there are some gaps in some data so things may not be strictly directly comparable. Put it like this, if this was a PhD thesis, I wouldn't want to be defending it. It's meant to be useful analytics, and augmenting your decisions, rather than making suggestions for your decisions. N.B. - Jancis scores are extrapolated using the formula 100_point = 60 + 2*20_point - it's crude but seems to work (17.5 -> 95)
Can I take a copy of the spreadsheet?
Yes, absolutely! You will only have read only rights to these, so if you want to analyze further, you'll need to take a copy.
Can I contribute?
Yes - just drop me an e-mail, or find me on Discord linked below, and I can add you as an editor, let's discuss first what you might want to do.
What is an R^2?
An R^2 measures, in effect, how well a prediction performs. We predict price is exponentially related to the score, and don't use any additional information (eg right vs left bank, white vs red vs sweet). By the end of en primeur, these scores will typically be about the 0.6 mark, which means we can account for about 60% of the price of a wine through an exponential of the score alone. Please note that this is not meant to be a predictive or descriptive model - eg trying to predict the 'correct' price for a certain wine. You could try to build that using features such as which bank/AOC the wine is from, the weather, whatever else. This model is only trying to work out, for a price, what's the best value for money you can get. If you could build a model that accurately predicted the correct prices, in principle you could 'trade' against that, eg you would expect actual price to converge with ideal price over time. I do not expect wines to increase or decrease price purely on the basis of whether they are above or below our predictive line on this model.
Can I get a referral code for a merchant?
Yes, absolutely, just shout. Both Lay&Wheeler and Berry Brothers & Rudd do them if you'd like one :)
How can I say thank you?
Shoot me an e-mail, but monetary donations are not necessary or needed. If you really want to spend some money to say thank you for this - please donate to a charity of your choice! :)
Why did you do this?
When it comes to wines being released, merchants have a way of putting specific spins on it to try to push the wines - at the end of the day, end consumers not buying wines don't pay the bills. This is all meant to be a strive to provide some basic analytics to help consumers on their buying journey, and make them a more informed customer. En primeur can be an amazing time to buy wines (2019, some of 2020 as examples), or it can also be an abysmal time to buy wine, where all you do is provide a negative interest rate loan to the Bordelais. I also just find it really interesting, and it's the closest I get to doing anything technical anymore.
Do I just buy the highest rating vintage?
No - you should try to find a critic who connects closely to your palate, and use their perspective. This score chart doesn't take into any consideration around styles of wines, alcohol, pH, or anything like that - and we all know scoring systems in wine are fundamentally a bit flawed at least.
Can I embed this data on my commercial website?
For personal use, you're welcome to do anything you like with this data, but please (where possible) try to inform me and let me know - if any cool ideas are coming up I'd rather incorporate them into the spreadsheets here. Please do not re-use these insights and analytics commercially without my expressed permission. Any re-hosting or use of this data by anyone should also include a link back to this website.
Reddit Community - https://www.reddit.com/r/WineEP
London Wine Meet Up Community - https://www.reddit.com/r/WineEPLondon/ (set to private)
Singapore Wine Meet Up Community - https://www.reddit.com/r/WineEPSG/ (set to private)
(Unaffiliated) - BordOverview is a really helpful link for looking at historic scores and prices -Click here
(Unaffiliated) - Wineberserkers -Wine-berserkers
You can e-mail me at admin at enprimeur.info