Reading the auction valuation badge on your Premium identifications

Premium identifications carry an auction-price-range badge alongside the predicted grade. It’s a useful, honest number — but only if you know what it is, what it isn’t, and why it comes as a range rather than a single figure.

Where the badge appears

Open any identification at /identifyon a Premium account. On the result page, immediately under the top candidate’s ruler and denomination, you’ll see two badges side-by-side:

• Predicted grade.P, F, VF, EF, or UNC — covered in detail in Reading the predicted grade.
• Auction range.A £-denominated range, e.g. £45–80, with a small “recent sales” subtitle showing the number of comparable lots used.

Click the auction badge and a panel opens with: the reference used to filter sales (e.g. North 1037, Spink 1395, S.1395), the date window of the comparables (typically the last 36 months), the auction houses included, and a small histogram of the actual hammer prices.

How the range is computed

We don’t make up prices and we don’t use a single catalogue figure. The range is built from real auction results, filtered down to comparable coins.

Sources

• Spink— quarterly numismatic auctions, the standard reference for hammered British silver.
• Noonans(formerly Dix Noonan Webb) — monthly coin and medal sales, strong on medieval and post-medieval English.
• Baldwin’s— specialist sales, including their joint sales with Stack’s Bowers.
• Heritage Auctions— for higher-grade Roman and Anglo-Saxon material that crosses the Atlantic.
• CNG(Classical Numismatic Group) — the dominant reference for Roman and Greek coinage at all grade bands.

Filtering

For each identification, the engine pulls comparables matching:

• Same reference number (North, Spink, BNJ, RIC for Roman).
• Same grade band (the predicted grade plus one band either side).
• Sales within the last 36 months, weighted toward the most recent 18.

The lowest 5th percentile and highest 95th percentile of the filtered set become the range. We trim the tails because auction results have outliers in both directions — a die-link rarity making 10× book, or a damaged piece making a fraction of one. The trimmed middle is what most coins of this identification, in this grade, currently realise.

Why a range, not a single number

Any single-number valuation for a detector find is dishonest. Even within a single grade band, identical references sell for wildly different sums depending on:

• Eye appeal.Two coins both VF on technical wear — one with a sharp portrait, one with a flat one — will make double-digit price differences.
• Toning and patination.Original cabinet tone adds 15–30%. Cleaned surfaces subtract more.
• Provenance.Pedigree from a named cabinet (Brand, Lockett, Slaney) adds a premium that can be larger than the coin’s technical value.
• Market mood. A strong sale earlier in the season lifts comparable lots that follow it. Coin markets are not efficient.

A range honestly represents the spread we can see in the data. A single number would pretend the spread doesn’t exist.

The grade-to-price relationship

Grade is the single biggest driver of price within a type, and the multipliers between adjacent grades are larger than people usually expect. As a rough rule for hammered British silver:

• P to F: 1.5–2×
• F to VF: 2–3×
• VF to EF: 3–5×
• EF to UNC: 2–4× (and often more for non-milled coins where UNC is almost theoretical)

Compound those across the scale and a UNC hammered silver penny can be 30× the value of the same type in F. This is why the predicted grade matters more than the predicted reference for the price you’ll actually realise.

Worked example — Edward I London Class 3g penny

North 1023, Class 3g, London mint. A common type, plenty of auction comparables. Here’s how the badge ranges out across grades, based on the 36-month comparables in our data:

• Predicted F: £15–25
• Predicted VF: £45–80
• Predicted EF: £120–200
• Predicted UNC (rare for this type): £350–600

Same coin, same mint, same class — the difference is wear. A modest grade improvement — the difference between a flat-portrait F and a sharper VF — can be the difference between £20 and £70 at auction.

What the badge is NOT for

Specifically:

• Not a retail price.Dealers retail 15–40% above auction hammer. If you’re buying rather than selling, expect the higher end.
• Not a probate value.Probate valuations are a regulated process by a chartered numismatist. Don’t submit our range as a probate figure.
• Not a fixed price.The badge updates as new comparables come in. A coin priced £45–80 today might be £60–110 in eighteen months if the market for that series strengthens.

Caution flags inside the panel

Click into the badge and you’ll occasionally see caution chips below the histogram. These flag situations where the range may not apply to your specific coin:

• Possible variety.The engine has spotted a legend or pellet pattern that could indicate a scarcer sub-class. A Class 3g and a Class 3g/4a mule look almost identical, but the mule is 5× the price.
• Possible die-link to a recorded sale.Coins that share dies with a published cabinet can carry a provenance premium. The engine flags candidate links but doesn’t claim them — you’d need an expert to confirm.
• Hoard-related discount possible. If the coin shows the surface signs of a known hoard (e.g. specific patination patterns) and that hoard had a large dispersal, comparable lots may already include hoard examples that softened the market.

How to read the range honestly

• Your coin is probably in the middle of the range. If you’re photographing well-lit detector finds without cherry-picking, expect the median, not the high.
• Eye appeal can push you up the range. Bold portrait, sharp legend, original surfaces — aim for the upper third.
• Cleaning pushes you down. If you’ve been heavy-handed with anything beyond water and a soft brush, expect the lower third or worse.
• The range moves. Don’t print today’s number and frame it. Re-check when you’re actually selling.

Next steps

• Reading the predicted grade — the input that drives the range.
• Insurance-grade PDF reports — turning identifications into named-collection cover.
• The reasoning chain demystified — how the engine arrives at the identification the range hangs off.