Collectible categories · screening model v1
Fifty-six collecting categories, sized by worldwide audience and market value, scored on the eight things that actually decide whether an object appreciates. Every row carries a grail you can't afford and an entry ticket you can — from a $20 Treasure Hunt to a running classic car.
The ranking is yours, not mine. Move the weights to say what you care about and every row re-sorts. The default weighting favours the thing that has historically mattered most: a young audience chasing an object nobody can manufacture more of.
| # |
|---|
No category clears those filters. Loosen the entry ticket.
Right is a faster-growing audience; up is a better fit to your weights. Bubble area is the estimated number of collectors worldwide — big bubbles are liquid but crowded, small ones are where mispricing survives. Hollow rings are categories trading at or just past a hype peak.
Almost every collectible boom runs on the same engine: people buy at 40–48 the objects they wanted at 8–16 and couldn't have. That lag is roughly 33 years, and it is the single most reliable timing signal in the whole category. Pick the year you'd want to be selling, and the clock tells you which childhood you should be buying now.
2032
Peak-spending cohort born 1989, aged 43 at sale.
Buy objects from
1997 – 2005
Buy while that cohort is 28–35 and still spending on rent, not relics.
Childhood windows and what they mint
Eight factors, and only three of them are about demand. The other five are about whether the market can support a price at all — which is where most amateur collecting money dies.
The most underweighted factor by far. Pokémon can print more Charizards; nobody can manufacture another 1994 Contax T2, another closed-mine specimen, or another run of a group-buy keycap. Categories where the issuer still controls supply cap out fast.
Is the median buyer getting younger or older? Stamps and vintage guitars have superb supply lock and are still poor bets, because the people who wanted them are exiting. Film cameras have identical supply lock and a buyer base that got 20 years younger.
The single biggest step-change in any collectible's price is the arrival of credible third-party grading. PSA did it to cards, CGC to comics, PCGS to coins. Categories with no grading standard trade at a permanent discount — and that discount is the opportunity if grading is plausibly coming.
Not "is it valuable" but "can I sell it Tuesday at a price I can look up". A graded card sells in an hour. A mineral specimen sells when the right person walks past your booth in Tucson. Both can be good; only one is an asset.
Where the category sits in its own hype cycle. Buying at peak has been the reliable way to lose money in every category on this list — sneakers 2021, rare plants 2021, sealed games 2021, whisky 2023. Cooling is often a better entry than climbing.
Both subtract. Counterfeits cap the price of ungraded material; carry cost is rent, insurance, humidity, hydrolysis and breakage. A sneaker's midsole crumbles. A cask evaporates. A plant dies. A card sits in a box for forty years and does nothing.
These are not hypothetical. Each one has a named corpse in the data above.
One honest note about the whole exercise: category CAGR is not your return. It measures how fast the money flowing into a hobby is growing, which drives the price of the scarce top end far more than it drives the price of the ordinary item you can afford. A 10% category CAGR with fixed supply and a young audience is a tailwind. It is not a yield, and nothing here is investment advice.
Market sizes and CAGRs are drawn from published 2025–26 industry research, listed below. Where firms disagree — and for collectibles they disagree a lot — the table carries the midpoint of the credible range and the detail panel names the spread.
Worldwide collector counts are estimates, not measurements. Nobody counts collectors. Each figure is derived from reported market size divided by a category-typical annual spend per active participant, then cross-checked against any published platform or player numbers. Confidence is marked per row: High where a real headcount exists (42M film shooters, 30M+ Pokémon players), Medium where the derivation is stable, Low where it's a reasoned guess. Treat them as orders of magnitude.
Scored 1–5 by judgement, not by formula — they encode market structure (does grading exist, can supply expand, who is buying) rather than anything measured. They are the opinionated part of this page. Disagree with one, move its weight to zero, and the ranking will tell you something different.
A weighted sum of the six positive factors plus the two negative ones inverted, normalised to 100. It has no units and no meaning outside this page. It is a way to sort a list, not a valuation.
Sources