Hold it or Drink it: What does the data say about aging wine?

Rachel Woods
Feb 12 · 8 min read

It’s said that only 1% of the wine produced in the world is meant to be aged. Now while that statistic isn’t well cited, and likely is skewed by bulk produced mass-market wine, it still brings the question: what does it even mean to age wine?

I thought this would be a challenging wine question to quantify. And I was right… you can’t just look at some quality rating of each vintage because you’re more likely to capture vintage-to-vintage differences, AND it’s usually unclear when a rating was given. You really need data about the same wine over time to control for a lot of variables.

Imagine a wine nerd, tasting at Tablas Creek in Paso Robles, stumbling upon these vintage charts that quantify the drinkability of wines at a given point in time. After learning that the team literally opens up bottles 3–4 times a year and rates their drinkability, and has been doing it since 2006, I knew this would be the perfect dataset for the task. Jason Haas, partner and general manager of Tablas Creek, was kind enough to share 32 of their past vintage charts and the analysis began.

So what did we learn?

The main things you need to know about aging wine are:

  • Wine evolves in the bottle over time
  • Not all wine is meant to be aged… there is such a thing as too much time
  • And similarly, some wines reach peak drinkability fairly early in their lifetime
  • Some wines go through a closed period
  • Not all people like drinking aged wine

While these are general concepts that you can read about in an abundance of blogs on the topic, I focused on confirming these with Tablas Creek wines using the dataset described above. The results are quite fascinating!

… let’s dig into each of these!

Wine evolves in the bottle over time

… and the evolution varies per wine. Using the vintage charts data, I was able to recreate an estimated timeline per bottle of when the wine was at different points of drinkability. When I grouped the wines by style (sticking to major categories — light white, full white, rose, light red, medium red, full red, and dessert), we can see that light whites and rose tend to hit peak maturity around 2 years after harvest. Whereas medium and full red wines hit peak maturity around the 6–7 year mark.

Another observation from above is how variant the time windows are even within wine categories. Jason pointed out that this is one of the challenges with many of the generalized “drink window” resources out there. While they can be a great rough guideline, each wine evolves at its own pace — making wine-specific drinkability ratings even more useful.

Not all wine is meant to be aged

By looking at a plot of a wine’s drinkability over time, we can see that most wines do tend to get better with age. However, rose and light whites experience a significantly different trend, with rose almost always ending up always past its prime around the 5 year mark. Again, this is based on Tablas Creek’s wines, but generally people seem to agree that lighter whites and roses are made to be drunk young. (Although you can certainly even age some rose).

Winemakers are your best ally here in trying to understand if a wine can or should be aged. First of all, many winemakers hold back their wines to release them when they think they are closer to drinkability — sometimes waiting 3 to 5 years to release the wines. Alternatively, you can just ask the winemaker: how do they expect the wines to evolve over time? They are the ones who have crafted the product, so they are really the best positioned to “predict” the lifetime of a wine.

We can also look at this question from the perspective of what attributes about a wine influence its ageability? I ran a linear regression to try to predict when each wine hit maturity measured by the years since the grapes were harvested. For our input variables, we are using an estimation of the characteristics of wines of this type (not specific to Tablas Creek). As you can see below, we have an R squared of 0.27, meaning that the regression model is able to explain 27% of what influences time to peak maturity using the simple input variables of body, tannins, sweetness, and acidity. Of course, one obvious improvement here would be to use the actual characteristics of the wine rather than a general wine type average, but it’s fascinating to see that these attributes are all statistically significant predictors of time to maturity.

Cool, so then can we model the entire lifetime of a wine? Maybe, but you need a MUCH more complex model. Taking an example from a blog Jason wrote on the closed phase:

“But a wine’s life cycle is not as simple as a linear curve, or even a bell curve, would suggest. Some wines — think most dry rosés — are at their best when they’re at their youngest, and fade relatively quickly. Most ageable wines actually have curves with two peaks: a youthful peak … and a mature peak some years later…”

So yeah… a much more complex model.

One last thing to note here is that it is expected that all wines do eventually fade. One limitation with the Tablas Creek data is that they only have vintage charts for wines going back to 2000. Many of their wines stand up to that amount of time, but wine is no different than anything else: all good things must come to an end.

And similarly, most wines are drinkable fairly early in their lifetime

Below is a graph of the average years to maturity since the grapes were harvested. Keep in mind that this is “since harvest” so when you open a 2017 Viognier in Spring 2019, it has been 1.5 years since harvest.

Many wineries will choose to make styles of wines that can both be opened early, but can also be held onto for a decade or more. This gives you the flexibility as a consumer to drink the wine when you want to drink it. Or even better, buy a case of it and enjoy it at different points in its lifetime!

Some wines go through a closed phase

What is the “closed phase”? Jason has a great explanation on his blog:

“… many wines have an intermediate stage where they are less enjoyable than they were, and less than they will be. This intermediate stage is often referred to as “closed” or “shut down”. You can equally think of them as teenagers: no longer children, with the charms of youth, but not fully adult either, often gangly and awkward, prone to moodiness and unpredictability.”

I took a closer look at the medium and full bodied red wines in our dataset to understand more about this “closed phase” phenomenon. Below is the distribution of the chances of a wine being closed given our dataset. This means that if you were to pick a medium/full bodied red off of the Tablas Creek wine list and it has been about 5.5 years since the grapes were harvested, there’s a 10% chance that it is in a closed phase. Helpful? Not really. In our opinion, the takeaway here is that it’s really hard to predict when a wine might be closed. Ideally you have something like these vintage charts to guide you. But if you don’t, and you open up a wine that’s 5 to 9 years old and it tastes off, don’t assume it’s past its prime and throw out the rest of the bottles in your cellar — decant the one you opened (it usually helps) and wait a few more years, hoping the wine will come out even better on the other side of its awkward years.

Not all people like drinking aged wine

So we’ve spent all this time talking and digging into data about how wines evolve over time. But it is important to remember that aging wine is a personal preference and there’s no “right or wrong” answer. We asked our followers on @the.wine.nerd and 21% of people said they don’t like drinking aged wine (n=136).

So what should you think about aging wines?

Aging wine doesn’t have to be complicated. Jason believes in making these vintage charts because Tablas Creeks’ wine program lends itself very well to benefitting from knowing how a wine is doing as it ages. Jason explained that Rhone varietals particularly have complicated life cycles and don’t always evolve in a linear path. And if there are ways that he can help their customers understand by giving them a tool like these vintage charts, then he should make them!

Don’t feel bad about not aging your wines. 40% of wine nerds said they don’t buy wine to age (n=136). Jason’s thoughts for people that don’t age the wines — “that’s great, we’ll keep making new ones!”

A good strategy for aging wines is to buy more (of the same) wine. Jason explained that he has made the decision in his own life to buy fewer wines, but MORE (eg. cases) of the wines that he loves so that he is able to enjoy and see wines evolve over time. For example, if you buy 1 case (12 bottles) of the same wine, you can open a bottle a year for the next decade!

Look for opportunities to try past vintages. This can help you understand if you like aged wine, and if the wines made by that producer age well. For example, Tablas Creek does a public 10 year retrospective tasting and has other opportunities to taste previous vintages via their vertical tastings.

I hope you learned a little bit about what it means to age wines; I know that I did through this research project! And special thanks to Jason for sharing his data and knowledge to support this work.

Rachel Woods

Written by

Data scientist, techy creative, wine nerd.

The Wine Nerd

Sharing research and insights about wine, for the consumer and the business, backed by data analysis and research.

More From Medium

More from The Wine Nerd

More from The Wine Nerd

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade