All Bordeaux Classified (ABC) - geeky, but very handy tool

Dear forum members,

In the end of 2018 I performed my first research of the Left Bank Bordeaux. I aggregated thousands of pro and user ratings in attempt to have an updated and weighted version of 1855 classification. I published an article with 20 top estates ranked according to the results of my study. Later using the same approach I studied the Right Bank Bordeaux.

However, that didn’t answer many questions I had remaining. Just to list a few of them:

  • Where do the more affordable classified growths stand in ranking?


  • How do the second labels/wines compare to classified growths? For example, Les Forts De Latour is on par with 2nd, 3rd, 4th or even 5th growths?


  • What level do wines from Graves compare to?


  • How about unclassified (like Sociando-Mallet), and Crus Bourgeois, and Exceptionnels? How do these wines compare to all of the above?


  • Let’s take numerous Pomerol wines and see how they compare to St. Emilions?


  • Are there any wines from satellite appellations which might be showing on par with neighbours from noble villages?


  • And what about many small bottlings and garagiste wines emerging?


  • Any decent negoce wines?


  • Why not compare Left and Right bank wines? Would Chateau De Valandraud rank closer to Pichon Longueville Comtesse De Lalande or Leoville Barton?

This all gets quite practical when I encounter with unfamiliar claret. Well, I have my own relative view on about 100 wines. But I want some benchmark to evaluate the unknown wine against, which I find in the restaurant wine list. It would be helpful to get a sense how this wine compares to other wines I am familiar with. Usually, I would open wine-searcher.com and look up the wine. But soon I realised there are over 1000 clarets that have 88-90 rating. So the number doesn’t really tell anything. It would me much more convenient to see how given wine relates to another.

We often stick to usual suspects. Wines we know or at least ring the bell. Something you find at your local importer, some forum recommendations, maybe wine magazine, etc. What could be another source to scout for something new? Take an appellation and find a new wine there. Not the trash one, but the one with good value for money spent.

While crunching numbers in order to find the answers to many such questions, I developed a handy tool for myself. It contains over 160K data points for 1350+ wines for all the vintages from 1981 to 2019. 30+ appellations, all existing classifications. I linked wines to their respective profile pages at wine-searcher.com and cellartracker.com

I have been playing with this data and developing this tool for the past three years. Don’t even ask me about the number of hours I have invested in this baby ))))

But it pays off, I must admit. I can now sort all the data the way I want and thus easily navigate through to satisfy my curiosity. So easy to look up, compare, discover or simply benchmark.

I keep updating the data all the time. Recently I have added average prices, number of user reviews for each wine on cellartracker, etc. I plan to identify organic, biodynamic wines in the future.

You might say it is self-promotion. Well, yes, but I don’t mind receiving some extra credits for immense work I have done.

Here is the link to a demo data set to give you the look and feel of my All Bordeaux Classified (ABC) tool. Check it out and let me know what you think of it. I added notes with detailed explanations to all the columns. It’s online Google Spreadsheet, so you can use a very powerful built-in functionality by sorting and slicing data the way you want.
All Bordeaux Classified.jpeg
Look forward to hear from you, folks!

3 Likes

Wow this is very cool and fun to look through. I can tell it must have taken an immense amount of work. Well done. I’m not clear what contributed to the final ranking, though. It’s not the overall scores alone, what else did you consider? For instance, what went into ranking Leoville Las Cases (avg score of 92.1) over Le Pin (avg score of 94.9)? Thanks.

This is great - and clearly the result of a crap ton of work. I think you did a good job synthesizing the data out there and would probably agree that this represents the relative prestige of the various chateaux/cuvees right now

First of all, thank you for your kind words!
That’s pretty much what I was trying to solve over the past years is how to come up with ranking that actually takes in account so many factors and remains balanced. Even though my university major is applied mathematics, I did consult with a few professors in math from the top universities to hear their point of view. And I must tell you it is tricky as it’s very difficult to reach the consensus. So my ‘ABC rating’ takes in account how each wine performs in each vintage. Vintages are different. Some of them are more competitive, others have overall high ratings due to good weather conditions, and you have disaster vintages. That’s why average rating doesn’t help much. Then you have wines which are rated in one vintage and not rated in another. The issue of data sufficiency is important. Or you face the challenge when you have many data points, but they are not consistent. Discrepancy between the ratings is huge. ‘ABC rating’ reflects the consistency and reputation of wines over many vintages.

Two avg. scores that you quoted for LLC and Le Pin are actually averages taken from cellatracker user rating, which vividly shows that user ratings are less reliable as opposed to the professional ones. Columns E and F in my chart have the data from cellartracker. Avg. rating and total number of reviews. You can see, for example, how popular are Cos and Leoville Comtesse.

If we take the above two wines LLC and Le Pin, have a look at the column D, which shows avg. pro ratings. There you see 93.81 and 93.77 scores respectively. Columns A, B, C take the level of geekiness of this tool to a very high level. Here you can see relative confidence index (how consistent are the ratings for the given wine), total number of data points and finally the number of vintages rated. Low confidence index and lack of data points are highlighted with red colors. Good index and sufficient data give you green. So when you play with data sets, you can adjust whether you want to only work with greens or you might want to broaden the range by including the reds and yellows.

I think the most interesting is analysing many other wines, not from the top50. This is where it gets really interesting and insightful. New discoveries, surprises, etc.

Pontet-Canet and Lynch-Bages have been buzzing for years now. These two 5th growths performing on par with 2nd/3rd growths. How about Chateau Grand Puy Lacoste from Pauillac. This 5th growth is a high performer, but somehow didn’t get the hype like the other two. My study shows it should. And especially given the price (all EUR): 123 and 144 for P-C and L-B, and rocking 79 for GPL.

I can talk on and on…

THank you so much for this acknowledgment. This means a lot to me.

I assure you, that the most exciting part is those other 1300 wines. You can slice and dice the chart to fish for great clarets under the radar. Rising stars, emerging micro wineries, etc.

Thank you for this tool. It’s clear that it took a lot of work. I’m pretty new to wine and I’ve been trying to find resources to learn as much as I can. I’ve been looking for a something like this to aggregate scores to help decide what wines to explore/purchase because manually going through CT and expert scores quickly becomes time consuming. I’ve already found some useful things in the spreadsheet to help guide future purchases.

Thank you, Lauren! Your support and kind words are much appreciated.
Would be great to learn about new use cases. Once you get yourself familiar with the tool, please share insights on how this tools helps you.
Even though I developed the tool, but more people start using it, the more I can learn from them and further enhance it.

Cheers!