Drinking windows explained: when to open the bottles you're saving

I moved countries with a bottle I had been saving. It went into the shipment carefully, came out carefully, sat waiting for an occasion good enough, and by the time I opened it the fruit had gone. What was left was structure and dried edges — perfectly sound, technically not faulty, and completely past the thing I had been saving it for.

Nobody had failed. It was stored well. The vintage was fine. I simply did not notice a window closing, because nothing was watching it and I was busy having a life. The bottle had been quietly ready for years, then quietly stopped being ready, and the only signal was one I could not read until the cork was out.

That is the actual problem with drinking windows. Not that they are hard to estimate — plenty of people will estimate one for you — but that an estimate nobody surfaces at the right moment is indistinguishable from not having one.

A window is usually written as a pair of years, and the pair is routinely misread. It is not “undrinkable, then good, then undrinkable”. Almost every wine is drinkable well outside its window. What changes is what you get.

• Before the window the wine is usually fine and often delicious, but the parts have not knitted. Tannin sits separate from fruit; oak sits on top rather than in. You are drinking the ingredients rather than the dish.
• Inside the window the primary fruit is still present and the secondary character has arrived. This is the overlap the window is trying to name — not a peak day, an overlap.
• After the window the wine is often still interesting and sometimes lovely, but the fruit has receded and what remains is structure, earth and dried notes. Some people prefer this. What you cannot do is get the fruit back.

So the honest reading of a window is: the years when the thing most people are saving this bottle for is most likely to be there.

There is a temptation to trust a precise window more than a vague one. It should be the reverse. “2028–2031” on a bottle nobody has tasted at maturity is false precision; “2027– 2035” is someone admitting how much they do not know.

The uncertainty is real and it comes from several directions at once: vintage conditions, how the producer made that particular year, your storage temperature and its stability, bottle variation, closure variation, and — not least — what you personally like. Someone who loves primary fruit and someone who loves tertiary development will disagree about the same bottle on the same evening, and both are right.

Treat a window as a probability spread, not an appointment. The useful question is never “is it ready?” but “how much am I risking by waiting another year?”

If a bottle arrives with no guidance at all, these are the crude defaults I start from and then adjust. They are starting points, not claims.

• Most whites and rosés: one to three years from vintage. The exceptions are real but they are exceptions — good Riesling, white Burgundy, Chenin and Hunter Semillon can run decades.
• Everyday reds: two to five years. Most wine sold is made to be drunk now and does not improve for being forgotten.
• Structured reds — classed-growth Bordeaux, Barolo, Northern Rhône, serious Rioja: five to twenty-plus, and the low end is often too low.
• Sweet and fortified: longest of all, and the ones most likely to outlive the person cellaring them.

Adjust downward for warm or fluctuating storage, and for half bottles — a 375ml ages appreciably faster than a 750ml of the same wine, and a magnum slower. Size is one of the few variables you can read off the shelf.

Every published window quietly assumes proper storage, and most wine is not stored properly. This is the single largest source of divergence between the estimate and your bottle — larger than vintage variation, larger than critic disagreement, larger than anything else you can control.

The direction is simple: warmer is faster. A cellar at 12°C and a cupboard at 20°C are not storing the same wine on the same schedule, and the gap compounds over a decade rather than adding up. A bottle kept in a warm flat can arrive at its window years early and leave it early too — which is exactly the failure I described at the top, since a bottle that travels has usually spent time somewhere warm.

Stability matters as much as the number. A constant 18°C treats wine better than a space that runs 12°C in February and 24°C in August, because the seasonal swing pumps the cork. If you are choosing between a cool spot that fluctuates and a warmer one that does not, take the steady one.

The practical consequence for windows: if your storage is warm or variable, shift the whole window earlier and narrow it. Do not treat a published 2030–2040 as applying to a bottle that has spent six years in a kitchen. It has already spent some of that window.

Here is what I got wrong for years. I thought the hard part was estimating the window. It is not. Estimates are cheap and mostly good enough. The hard part is that a window sitting in a record is silent, and a silent record cannot compete with the six bottles visible at the front of the rack.

What you need is not a better estimate. It is for the cellar to raise its hand: to say these four are in their window now, and this one leaves it this year — at a moment when you are choosing what to open, not when you are doing admin.

That is what drinking windows are for in CellarPals: every bottle carries a from-year and a to-year, the cellar can be filtered to what is ready, and the ones closing get surfaced rather than waited on. The estimate is the easy half. The surfacing is the half that saves a bottle.

Three habits, none of them demanding:

• Set a window when you enter the bottle, even a lazy one. A wrong window you can revise beats an empty field you will never fill. Ten seconds of “call it 2029–2034” is worth more than the perfect answer you were going to look up later.
• Check what is closing before you buy more. The cheapest way to stop losing bottles is to notice you already own something that needs opening this year.
• Open one early on purpose. If you have several of something, drink one before you think it is ready. That single data point about your bottles, in your storage, to your taste beats any published window — and it is how you learn to revise the rest.

Where the bottles live matters here too: a list of what is ready is only actionable if it can tell you which shelf to walk to. That is why the window and the cellar map are the same feature in practice, and it is what the cellar layout guide is about.

A window from a critic describes their bottle in their storage against their palate. Yours is the only one that matters, and you can measure it — slowly, cheaply, and with genuine pleasure — if you own more than one of anything.

Buy in threes where you can afford to. Open the first early, the second when you think it is ready, the third when the record says the window is closing. Write down what you found each time. After the second bottle you know more about that wine in your house than any published note can tell you, and the correction generalises: if your bottles are consistently ahead of their windows, they all are, and you should be shifting every estimate in the cellar.

The alternative — buying one bottle of many different things and hoping to catch each at the right moment — is how you end up with a cellar full of single chances and no way to learn from missing one. Depth is not just for drinking; it is the only feedback loop a cellar has.

Revise the record when you learn something. A window you have adjusted twice from experience is worth more than any number you copied in, and the adjusting is what turns a cellar from a store into a practice.

Open it anyway, and open it attentively. A bottle past its window is the most informative thing in the cellar: it tells you what your storage does over that many years, and it recalibrates every sibling bottle you still own.

And decant it less, not more. Wine that has lost its fruit rarely benefits from air — what is left is fragile, and half an hour in a decanter can finish the job the years started.

One closing caution, because a cellar organised entirely around windows becomes a schedule, and a schedule is a poor reason to open a bottle.

The windows exist so that good bottles do not die quietly at the back of a rack. They do not exist to tell you what to drink on a Wednesday. Plenty of the best bottles I have opened were opened early, on purpose, because the evening deserved it — and a wine drunk slightly young among people you like is a better outcome than the same wine drunk perfectly alone because the record said this was the year.

Use the window to avoid loss, not to schedule pleasure. That is the whole of it.

Everything above is how CellarPals works, because it is the cellar I keep. Fourteen days free, then $12.99 a month or $119.99 a year.

If you are moving from somewhere else, the CellarTracker comparison is honest about what carries over and what does not.

How to scan wine labels with your phone — And keep the cellar in sync.

All guides