Heat

Heat is how busy a set is — measured on the chase cards we track from it, not on every card it printed (50 per set is typical, from 12 to 154, 3,247 cards in all). Of those, it is the share of everything listed that actually sells on an average day. Bulk commons are left out on purpose: they trade in enormous volume at a few cents and would bury the reading under penny churn, while the chase cards are where demand actually turns up. That is an activity measure, not a value one: a set can sell briskly and still fall in price. The colours run cool to busy, and mean nothing about good or bad.

Sets get quieter as they age — a six-month-old set sells far more of what is listed each day than a five-year-old one does. So ranking sets on the plain number mostly ranks them by release date (rank correlation −0.86 against age). The age curve is refit every build across all 61 sets here, and the “for its age” figure is each set against what its own age predicts. The Overall view asks the other question: how busy each set is, plainly, with no age step at all. Bars there are directly comparable, the order is simply busiest first, and age gets its own column — which is how an older set sitting high among much newer ones becomes visible. Each row still carries its band, so a set can be the busiest on the board and still only ordinary for its own age.

Five regions cut from how far sets normally sit from the curve, not a score — named on a temperature, hot through quiet. “Normal” is the middle of the pack for a set that age; “running hot” and “running quiet” are the tails, and every one of the five is a reading for the set’s OWN age, not against the board. The curve explains about 81% of the cross-set spread today, but that figure leans hard on the few youngest sets — on sets over 2 years old alone it is 59%. A set sitting just over a boundary is not meaningfully different from one just under it.

Trainer Galleries, Shiny Vaults and Celebrations: Classic Collection are carve-outs of a parent set that is also on the board, so the two share cards. They are always shown — a measured row should not be optional. A carve-out has no logo of its own, so it borrows its parent's; the small "Sub" tag beside the set name is what tells the two rows apart. All of them currently sit above their age curve: that may be how chase-only rosters trade, or an effect of pooling a small card count. We have not separated the two.

Not a price call and not a forecast. It describes the last 45 days of trading and says nothing about the next ones. Sets with too few tracked cards or too few recorded sales are not on this board at all — absence here means unmeasured, never quiet.

We built a movement detector and tested it against a phase-randomized surrogate — the same series with its timing scrambled but its statistical shape kept. It fired no more often on the real data than on the surrogate. That refutes this detector on this series, not the idea of demand episodes: the surrogate preserves week-to-week persistence by construction, and our own residual autocorrelates negatively, so moves here bounce rather than carry. No badge, no episode timeline. Whether elevated pressure precedes price moves is untested, not disproven. Full figures and the script that produces them are in the methodology page.

A 45-day window on one venue (TCGplayer sold over TCGplayer supply), across 3,247 tracked chase cards in 61 sets — the chase rarities, never every card printed. Earlier supply history was a backfilled placeholder and is excluded, and a leading run of frozen values is trimmed off each card. Every window ends 3 days back because sales publish late. The listing count is unchanged day over day about 74% of the time, so the number of cards listed is a coarse series. Below 25 sales in 7 days no change figure is shown at all. Full disclosure, including what is dropped and when this board goes dark, is on the methodology page.