How to read the index

Every term we use to publish a value, defined in plain language. If a number on this site does not make sense, the word behind it is probably here.

The vocabulary

Nine terms cover almost everything on an artist profile or an index page.

Verified: A verified artist has enough auction history for the model to fit a price. Their work has traded often enough, and recently enough, that past sales say something about the next one. Verified artists get a value range with a confidence interval and a liquidity score.; Related: confidence interval, liquidity score, never blended.
Emerging: An emerging artist has little or no auction record. There is not enough sale data to model a price, so we do not publish one. We publish a momentum score instead. An artist moves from emerging to verified when the sales record catches up, not on our say-so.; Related: momentum score, never blended.
Confidence interval: The band around a verified value where we think a real sale would land. A narrow interval means recent sales agree with each other. A wide interval means they scatter, or there are only a few of them. A wide interval is us being honest about thin data, not us guessing.
Hedonic model: The regression behind every verified value. It prices a work from its measurable traits: the artist, the medium, the size, the sale venue, and the period, fit against every sale we have on record. The approach is standard in property and art economics. Break a thing into its features, learn what each feature is worth, add them back up.
Liquidity score: How often a verified artist's work actually comes up for sale. A high score means many lots a year, so a value is easier to act on. A low score means you may wait a long time for the right work, or the right buyer. It measures trading frequency only, nothing about quality or direction.
Momentum score: The single figure we publish for emerging artists. It reads exhibition frequency, gallery representation, and where attention seems to be moving. It is not a price and does not convert to one. We use a different word on purpose: the data can show direction, but it cannot carry a valuation.
Gallery asking range: What galleries are currently asking for an artist's available work, read from their own online shops. It is the seller's price, not a buyer's, so it is not a valuation and never feeds one. We show editions separately where the gallery gives an edition size, drop anything sold, withdrawn or no longer listed, and show the middle half of prices once there are four or more works, so one outlier does not stretch the range.
Never blended: Verified values and momentum scores stay apart. They come from different data and answer different questions, so we never average them, and no index mixes the two. Each artist sits on one track. A value range is all verified data. A momentum score means there was no price to publish.
Blend index and hedonic index: Two ways we roll individual artists up into one market number. A hedonic index tracks the modelled value of a fixed, representative basket of work, so it separates price change from a change in what happened to sell. A blend index combines two things: a like-for-like auction price level, measured by comparing each artist's sales with their own earlier sales, and the share of auction lots that actually sold over the past year, compared house by house. Neither one ever pulls in a momentum score.
Base value and rebasing: An index is a ratio, so it needs a starting point. We set the base value to 100 at a chosen start date, and every later reading is measured against that. Rebasing moves the reference point, usually to a more recent date, so current readings sit near 100 and are easier to compare. Rebasing changes the scale, not the underlying trend.

Want the model itself, not just the vocabulary?

Read the method