Wine Price Methodology: How We Calculate Every Number

Our wine price methodology separates each kind of market evidence before calculating anything. Most wine price pages show you one number and let you assume it means something. It usually does not. This page explains exactly what our numbers are, so you can decide how much to trust them, and catch us when we get something wrong.

What kind of number is this?

Every price on this site carries a label saying what it is, because these are not interchangeable and treating them as one is the single most common way buyers overpay. An ask is what a seller wants today. Nobody has agreed to it. It is the weakest form of price evidence, and it is also the number most easily found, which is why it dominates search results. A bid is a live offer at auction, a floor rather than a clearing price. A hammer is what a lot actually sold for. It is the strongest evidence of what the market will bear, but it excludes the buyer's premium and everything after it. An invoice is the total actually paid with fees included, and it is the rarest and most useful of the four. An index value is a published market level, not something you can transact at.

When you see a figure on this site, the label beside it tells you which of those it is, along with the venue and the date we retrieved it. If a number appears without that, it is a bug. Tell us and we will fix it.

Why do you separate retail asks from auction clears?

Because the gap between them is the actual information. Across the wines we track, retail asks typically sit meaningfully above auction hammer prices, and that spread is not a mispricing waiting to be arbitraged. It is the price of certainty. A retail bottle comes with known condition, immediate availability and a returns path. An auction lot comes with a lower headline and a longer list of things that can be wrong with it.

Averaging the two into one market price destroys that information and produces a number that describes neither transaction. So we plot them as separate series, and we never blend them into a single headline figure.

How far back does the price history go?

Up to 50 years for the wines with the deepest auction records, and every observation we hold is dated and attributed to the venue it came from. Coverage is not uniform. A first growth with a long auction history has far more evidence behind it than a small grower who first appeared at auction five years ago, and the chart shows you the thin patches rather than smoothing over them.

How is the score calculated?

The score out of 100 has two parts, weighted deliberately.

Critic ratings account for 40%. These are the published scores from professional critics, normalised onto one scale so that a 17.5 out of 20 and a 93 out of 100 can sit in the same average. Critics taste young wine in controlled conditions, which makes them good at judging quality and structure, and less good at telling you what a bottle is like on a Tuesday twelve years later.

Drinker ratings account for 60%. These come from drinkers rating bottles they actually opened at home, across many vintages and many storage histories. There are far more of these, they are spread across the whole life of the wine, and they are the better guide to what you will experience.

We deliberately exclude one input that other sites lean on heavily: aggregate scores derived from merchant listing data. A listing score measures how a bottle is being sold, not how it tastes, and folding it into a quality number quietly turns marketing into a rating.

Where a wine has too few ratings on either side, we show the parts we have and no combined score. A confident-sounding number built on four opinions is worse than no number.

What about fees, duty and shipping?

The price you see is the price at the venue. What you actually pay is higher, and by how much depends on where the lot is, where you are, and how the house charges.

Our landed-cost calculator works out the additions line by line: buyer's premium charged in tiers rather than at a flat top rate, VAT on the premium where the venue applies it, lot and payment fees, excise duty by volume, import duty, packing, shipping, transit insurance, storage, VAT on the delivered value, and the currency spread over the mid-market rate.

Each line shows its own explanation and links to its source. That explanation is the product. A model that says £1,480 without showing how is asking for trust it has not earned.

What we deliberately do not do

We do not fold condition and provenance risk into the price. A bottle with an unclear storage history is genuinely worth less than the same label with a documented chain, but mixing a probabilistic risk estimate into a currency total produces a number nobody can defend. We score it separately and show it separately.

We do not present a retail ask as realisable value. If you are thinking about selling, the number that matters is what you would net after commission, and that is a different calculation entirely.

We do not publish a page for a wine we do not have enough evidence about. An entity with no market observations returns a 404 rather than a page saying no data available, because an empty page holding a keyword is a doorway page whatever the intention behind it.

We are not paid to rank a merchant, a house or a wine. Nobody can buy a position on this site.

How do you handle corrections?

Our observation store is append-only. We never overwrite a price, we add a new one. That means every chart can be reconstructed exactly as it stood on any past date, and a correction is visible as a correction rather than a silent edit.

If you find a wrong number, tell us. We will fix it, log what changed, and say so on the page.

What identifiers do you use?

Everything is keyed to LWIN, the free Liv-ex wine identification standard. LWIN7 identifies a wine, LWIN11 adds the vintage, LWIN18 adds the format and pack size. Using an open identifier rather than one we invented means our data can be joined against anyone else's, which is the point of using it.