Methodology

1. The formula

The whole index reduces to one line, applied to every coin we track and summed:

Three terms, each one a fact we can point at:

• Lowest current dealer ask — for each coin, the cheapest price a dealer is currently asking for one in that certified grade. It is a floor offer, not a sale: an ask, the lowest one we can observe right now.
• Certified population — how many examples PCGS and NGC report having certified at that exact grade. This is the “shares outstanding” of the coin: the total graded supply, taken from the most recent census.
• Σ — we multiply ask × population for each indexed coin and add them all up. That sum is the index. The chart on the market-cap page is this same calculation run for each day in the window.

2. What “HPG” means

A coin qualifies as High Population Grade when at least 10,000 examples have been certified at the same grade by the major grading services (PCGS / NGC). That threshold is the whole point of the category. A unique coin has no market price — only a last hammer — because there is nothing to compare it to. Once tens of thousands of graded, condition-verified, interchangeable specimens of the same coin exist, any one of them substitutes for any other. That is what makes a single quoted price, a ticker, and a market cap meaningful at all.

These coins are not rare. They are common enough to be fungible-enough-to-price — the blue chips of numismatics, steady and known. They are still collectibles, trading above their metal content for design, history, and grade; they are not bullion. For the fuller picture of what HPG is and why it behaves the way it does, see Learn.

3. Data sources

Every number in the index traces to a dated observation in colcur's price-store. There are three inputs:

• Prices — dealer asking prices from CollectorsCorner, collected continuously by colcur's own scraper and written into the price-store. For each coin we take the lowest recent CollectorsCorner buy-it-now ask, in USD.
• Population — certified-population snapshots from the PCGS and NGC census reports, per coin and grade, dated. We use the most recent snapshot for each coin.
• Bullion floor — LBMA fixings for gold and silver, which set the metal-value floor under each coin and drive the bullion toggle (below).

Every coin is keyed by its stable colcur catalog identifier, so a price observation, a population snapshot, and a metal floor all attach to the same coin without guessing. The index is read live from the price-store: as more dealer asks and census snapshots flow in, more coins gain a price and population, and the figure fills in. A coin with no priced listing or no census yet simply contributes nothing until its data lands — the number reflects what we can actually observe today, not a backfilled estimate.

4. The honest caveats

This is the section that matters most. Read it before you quote the number. None of these are defects we are hiding — they are the known, named properties of the construction, and stating them plainly is what makes the figure citable instead of attackable.

• (a) It is an asking-price notional, not a clearing value. The index is the lowest current ask × the full certified population. No one could realize that figure by selling — it assumes every coin ever graded is worth today's cheapest listing. This is exactly the simplification an equity market cap makes: last traded price × all shares outstanding, including the shares that never trade. It is a standard, defensible construction — but it is a notional, never “what the coins would fetch if sold.”
• (b) Population is not free float. Most certified coins sit in collections and will never come to market this decade; they are permanently locked supply, the coin equivalent of founder stock. On top of that, dealers crack coins out of their holders and resubmit them chasing a higher grade, so a single physical coin can be double-counted across grades in the census. Both effects push the population — and therefore the index — up.
• (c) Prices are dealer asks, and can be thin. A coin's price is whatever a dealer is currently asking, and for some coins that rests on very few active listings. A single relisting can move a ticker. We show the source and the observation date on every coin so you can judge the depth behind a price yourself.
• (d) The source you pick changes the answer by a multiple, not a margin. A controlled, same-coin comparison in our price-store shows published price-guide quotes running about 2.2× the price the same coins actually realize on eBay. Which layer of the distribution chain you price from — realized sale, wholesale bid, retail guide, dealer ask — swings the valuation by multiples. We deliberately use the lowest dealer ask, which sits near the top of that stack, and we say so here rather than let the choice pass unstated.
• (e) The bullion toggle. Some HPG coins are modern bullion-class issues (American Silver Eagles, American Gold Eagles, and similar) whose value is dominated by their metal content rather than a numismatic premium. They are certified in enormous numbers, so they can swing the total. The market-cap view lets you include or exclude bullion-class coins; we default to showing both so neither view is hidden.

5. Could it be more precise?

Yes — and the path to a more rigorous index is well understood, because it mirrors what equity-index providers already standardized:

• Free-float weighting — replace the full certified population with an estimate of the supply that actually trades, proxied from observed annual turnover. This directly addresses caveat (b).
• A realized-comp line — publish a parallel series priced on realized sales alongside the lowest-ask series. The two bracket true value, and the gap between them is the liquidity discount from caveat (d).
• Survivorship adjustment — net out the crack-out-and-resubmit double-counting in the census so one physical coin is not counted in two grades.

That work is downstream, and we will add it. For now we publish the transparent, reproducible version: one public formula, the same data producing the same number every time. That reproducibility is the point. A number that improves in the open — with every refinement stated — is more trustworthy than a polished one whose internals you have to take on faith.

6. What colcur is not

colcur is a data and media property: it measures and reports a price. To be unambiguous about the boundaries:

• Not an exchange. The “like stocks” framing is a metaphor for how a transparent price makes a coin legible at a glance. It stays a metaphor.
• No custody, clearing, or execution. colcur does not hold your coins, match buyers and sellers, or settle trades. Coins stay in their owner's possession; we link out to dealers.
• Not investment advice. The index is a measurement, not a recommendation. Nothing here is a forecast, a guarantee, or a suggestion to buy or sell.