What an auction result actually says

Ask what an antique is worth and the honest answer is a question back: worth to whom, on what day, sold how? The nearest thing to a real answer is what objects like it have recently sold for at auction — a completed transaction between a willing buyer and a willing seller, in public, with a date on it.

That is genuinely good evidence. It is also routinely misread, because a single lot generates three numbers and they are not interchangeable.

The three numbers

The estimate is a range printed in the catalogue before the sale. It is set by the house, partly to guide bidders and partly to attract them, and it is frequently pitched low to encourage competition. It is a marketing figure. Nobody has paid it. Lots sell at multiples of their estimate and lots go unsold below it.

The hammer price is the winning bid — the moment the auctioneer stops. This is the number that most closely represents what the object was worth to the market that day.

The buyer’s premium is a commission the house charges the buyer on top of the hammer price. It is commonly 25% or more at the major houses, often on a sliding scale, and it may have tax on it as well. The house publishes its rates in its terms of sale.

So a lot that hammers at £2,600 may be reported by the house as “Sold for £3,328 inc. premium”. Both figures are correct. They describe different things.

Which one is evidence of what

If you are asking what would I get if I sold one, the hammer price is the closer answer — and even then it is before the seller’s commission, which is a separate charge again.

If you are asking what would it cost me to buy one, the premium-inclusive figure is the closer answer, because that is what leaves the buyer’s account.

The error to avoid is mixing them: comparing a shop’s asking price against a hammer price makes the shop look expensive by roughly the premium, and comparing a hammer price against a premium-inclusive one makes an object look as though it has appreciated when nothing has happened at all.

Patine shows hammer prices in its comparables and says so on the row, which is why a lot may read lower here than on the auction house’s own page. The link goes to that page so the difference is visible rather than hidden.

Estimate ≠ value, and unsold ≠ worthless

Two related traps.

An estimate quoted as a valuation is a claim nobody has tested. If a description says “one sold at Christie’s with an estimate of £5,000”, the sentence has told you nothing about what anyone paid.

A bought-in lot — one that failed to reach its reserve — is not proof the object is worthless. It may have been badly catalogued, wrongly estimated, or in the wrong sale. It is evidence of one failed attempt, no more.

Why asking prices are a different category

A live listing is not a sale. It is one seller’s opinion, and it can sit at that price indefinitely without anyone agreeing. Aggregating asking prices produces a number that looks like a market and is not one.

This matters more than it used to, because sold-price data is getting harder to reach. eBay closed its sold-item API to new applicants years ago and has since put sold listings behind a login; the large commercial archives have no public interface at all. The result is that a lot of what is presented online as “what it sold for” is in fact what somebody is asking.

The practical defence is to ask of any figure you are shown: who paid this, when, and where can I see it? If those three questions have answers, it is evidence. If they do not, it is an opinion, and it may be a good one, but it should not be doing the work of a fact.