I hope I’m not breaching forum rules by referring you to our new vintage chart - launched today in time for Christmas.
https://www.wine-searcher.com/vintage-chart
We’d love some feedback so we can improve it.
Thanks!
2 Likes
Hi Jules, love W-S! The aggregate vintage chart, not so much. I’m not sure if it’s worth your time when it’s showing 92-94 rating for fine French red for like 25 years. Of course you could say that about any of the aggregate scores. Maybe someone out there appreciates it, I just don’t get much value from it.
Pretty interesting! It seems like the rating system+color makes it just a bit convoluted to me. When looking at a chart like this I think I might prefer some sort of color with a corresponding legend, and maybe a resource to look at actual vintage data or something similar when clicking on the aggregated score. I think we’ve all realized at this point that ‘score’ is completely arbitrary.
Appreciate W-S putting this together.
As a high level tool for most wine drinkers (not necessarily on this site) it could certainly be a hand reference guide. Fun to look way back in time for me either way
Julian
Thanks for posting, and for asking for feedback. I only drilled down on the last decade of champagne vintages. I suspect you’re using critics’ ratings as a gauge. However, when sorting and winemaking skills can produce good wines in a challenging vintage, it is still not a good vintage. 2010 was difficult in Champagne, and 2011 even more so. They’re definitely not 93 point vintages.
Cheers,
Warren
Seems pretty useless, honestly. Having France and Italy as one region for the entire country isn’t helpful, nor are scores that rarely go below 90 no matter how poor the vintage really doesn’t make sense. Let’s see some 50s, 60s, and 70s for vintages that deserve it.
If you click the country it will drill down to finer-grained regions
1 Like
I really like the idea and I always appreciate additional data points, but I wonder how reliable/useful the methodology is, especially when it says things like 2018 was a better vintage than 2019 in Mosel or that 2019 was the worst vintage in Napa in the last 3 decades.
Who would have guessed 1913 Australia was a 100 pointer?
I will be sure to seek them out.
Hey there Julian! Welcome and thanks for the new chart…!
One thought: consider experimenting some more with different colors/hues for the vintage scores (deep red seems to correspond with the best vintage - but is this somewhat contrary to most cultural conventions?). Also, in theory if you’re going to explore the idea of different tones representing higher or lower scores, they should be a bit more distinctive so you can immediately distinguish years of quality. There’s definitely a fine line there - you don’t want it to be a kaleidoscope of colors and tones, but it seems a bit too compressed in terms of color/saturation in the point ranges that really matter (say 90-94 and 95+). From my view, if you have a 98+ vintage, it should quite distinctive so as to say “vintage of a lifetime!”.
I do see that there’s a bit more range as you get into the 80s, but in my opinion, you could make everything under 88 points the same color. At this point, the differences between an 85 and 88 range is not particularly useful (especially for buyers). Conversely, the differences between a 94 and 97 vintage are much more significant and ideally the colors would represent more contrast here.
All the best and thanks again!
P.S. One other thought, and this may be more controversial… point inflation is definitely a thing, and it might be useful to apply some type of modifier to match some of the point inflation over the decades. It would be hard to believe for example that 1990 is an equal vintage to 2011 in Bordeaux. (both 93)
I think this is a great idea, but putting California into one region is pretty ridiculous. We have such a wide range of climates that northern in Southern California could be separated and even with the mat subweedens could be separated. I’ll give you a great example. 2020 was a real challenge for much of northern California but Santa Barbara County had no major issues. And that’s just one data point and one example.
Finally some Jura info… but then I see 2017 given a 95 score… I love W-S. But it does not seem very helpful.
You can click on California to expand and see the sub-regions. There is actually more granualrity than in any other vintage chart I’ve ever seen.
Ooops . . . my bad . . .
Thanks for the correction.
Cheers!
I like that the vintage chart covers more regions and sub regions than most (or any) other vintage chart out there and that it goes back many more decades than most (but some specialized) sources. It would be interesting to know on what data you base your date on.
I love W-S and find it one of the most useful tools on the internet, but I think the scoring system here is problematic. For example, 2007 and 2013 are shown as among the lowest scoring Bordeaux vintages in the last two decades in the chart, which I think is directionally correct, but I think most Bordeaux drinkers would take issue with labeling those as 92 point vintages. I assume it has to do with the underlying database - perhaps lower end wines in those vintages are not being as actively bought/sold and therefore not registering? Is there a way to adjust for that, or tweak the scoring system to force more distribution between high and low?
Thank you for your suggestions – please keep them coming. We are keen to do the best we can.
I’m going to ramble on a bit here – but first you should know my background. I’m a maths/stats/computer geek – not a wine expert.
Yes – the colour scheme needs work! Is green good, or is good red hot?
We should probably make it clearer that you can drill down the regions.
Yes – we intend to provide you with more data – to allow you to see which wines/scores made up the average and compare average prices, etc. We also want to provide a way to click through and see available stores/offers for those wines.
Yes – the aggregation/averaging of scores does visibly remove the highs and lows. This is both good and bad. There are almost no 100-point average scores – but why would there be – was there a vintage where every critic agreed all the wines from a region were 100? The way I take that is to read the numbers as 95 is not 100th better but way better than 94 – because to move the average by just a single point across so many wines and many scores for a single region/vintage is significant.
If you are a wine expert – you probably don’t need a vintage chart. You probably know what wines you like and why, and whether you think Parker is a god, or not. Also – if you are a wine expert then you probably don’t want to think of Bordeaux generically – you think about specific wines and specific vintages. You know that you prefer Margaux to Petrus or whatever.
I quite like Sauternes. I find it disappointing that the acclaimed 2001 Yquem is not really that visible in the data – 2001 doesn’t stand out:Sauternes - Fine White Wines | Wine Vintage Chart
and not all critics agreed even for that one wine/vintage:
Where to buy 2001 Chateau d'Yquem, Sauternes | prices & local stores in UK
but it is a higher than average vintage for Sauternes.
2001 Wine Vintage Report, Bordeaux | Vintage Wines & Spirits
Scores/ratings are for the rest of us – who look at a wine list and need help – and don’t want to look like idiots when we order the “wrong” vintage. If you want to compare two specific wines in a restaurant perhaps – then look up those wines directly – a vintage chart won’t help.
In our research we found that it is not really fair to compare everyday wines with fine wines for a given region/vintage. So we force you to choose a price range. That’s not perfect – but it is a way of breaking down the data – maybe we should go further with that?
The next factor is that we are averaging across wines from a given region. If that region is the whole of France – then there are likely thousands or tens of thousands of scores being used and the averages are likely not that helpful. So we allow you to select from 3 price (quality) categories and also choose the wine colour/style. However – better still to drill down the regions from France to Bordeaux to Medoc, etc.
IMHO – user ratings are almost worthless, though Google likes them. The reason is that the usual 5-star rating has no common reference. One person might give 5 stars because they have tasted Petrus and Margaux and frankly they weren’t as good as this Latour. Other users might give Latour 3 stars because they think it’s not great value for money despite being delicious, or because it didn’t go with the salmon, or because they had an argument with their partner. Is it an absolute reference of taste or a relative value for money score? My palate is unrefined and untrained – but my user score is just a meaningful as anyone else’s. Enough consumer reviews of anything will average 3.6 (because maths/psychology) – how is that helpful?
I would like to think that wine critics (in fact professional reviewers of anything) are more objective than consumers and taste enough wine to have a more stable reference and so when they say 94 and not 98 they are more likely to be objective or tend towards it. Of course there is subjectivity included. Parker preferred a particular style, and some critics have a reputation for giving higher scores than others. When I look at some of the data I see a tendency to scoring higher in more recent years – is that a trend of better winemaking, or a trend in commercial the pressure to score higher?
Regarding the numbers themselves. Firstly there is the 100-point scale. We map each critic into the “accepted” 100-point scale. This is not straight-forward as nobody actually uses the whole range – most critics score between about 75 and 100. Jancis scores between 12 and 20 (or effectively between 15 and 20). We are trying to read precision from something which is not as precise as we would like. We could perhaps show the highs and lows of the scores as well as the average – or include the standard deviation, or median, even the mode! There are statistical techniques for managing this type of data – removing outliers, using the Bayesian average, etc. Most methods have benefits and problems.
We could also be selective about which critics/scores we include – but I think this is a dangerous road to follow. We have quality standards and already select critics carefully – should we go further and only build the chart from a subset of critics. To be global we need to embrace perhaps lesser known critics who specialise on particular less well-known regions.
More feedback, please.
3 Likes
Hmm very interesting. And thank you for being open about taking feedback here!
Quick one - how do I drill down on an iPhone (Safari)?
Jules, it might not fit a data driven model, but I don’t think there would be anything wrong with using a more subjective approach to vintage ratings. I’m not sure how you curate that and put a number to it, but I actually think critics (and enthusiasts) tend to be far more objective about vintages in retrospective and assessing a vintage 5 or 10 years down the line tends to be less noisy than trying to aggregate scores from barrel, a year out of the bottle or what have you.
The value of user reviews (say on Cellartracker) is that wines are being scored over a much larger period of time - whereas most critic reviews are before or right after bottling when wines are being released. This of course makes a lot of sense because consumers are thinking about wines that few people have tasted and need to make an informed decision about buying. In this sense, these types of professional critic scores are biased towards first impressions which are not always accurate - and depending on what wines scores are being considered in the data it may not give an accurate sense of the vintage. (For example, First Growth Bordeaux will likely distort the overall picture of a vintage). You might consider selecting a number of bellwether producers from each region that you use as a basis for the scores (and I would exclude First Growth and comparable wines from that list). And lastly, you might also consider using user reviews (like Cellartracker) to temper the data for older vintages weighting the scores more towards CT (or using a rolling average with these in the mix).
I realize that user reviews are somewhat sketchy to use, but in aggregate I think that CT more often than not captures the quality of a vintage for producers quite well. For example, if you go along multiple vintages of Leoville Barton you can capture a good sense of vintage to vintage quality. Critic ratings for the 2013 Leoville Barton were between 90 and 95. CT is sitting at 89.9 after 35 reviews - which seems more accurate as most critics have revised their scores lower over time.
Just click on the country/region name - if it’s in blue text then sub-regions are available.