Every market that has gone up for a year and a half produces the same video, and the video is usually right eventually and wrong for a long time first. We cannot tell you when. What we can do is what nobody making the video does: put the question to a tape we built ourselves, name the readings that would have to move for the answer to become yes, and print where they sit today. Eight of them follow, each drawn from a study we have already published, with the current value and the value that would count as a crash signal. Then the caveat, which is not small.
Eighteen up months, no 2% drawdown, decelerating
moonstone Index, month-over-month change
The index closed at 235.25 on September 14, its high, up 135% since it started at 100 on March 17, 2025. It has risen in every one of its eighteen calendar months, the weakest being February 2026 at +0.7% and the strongest March 2026 at +12.4%, and its deepest fall from a running high in 354 daily closes is 1.8%, in late October 2025. Over the twelve months to September 14 it was up 83.5% against 15.7% for the S&P 500, and over three months 16.0% against 1.4%; both readings roll daily, and one day later they read 81.2% and 14.7%. The one thing in the level that a bear can point at is the slope: the spring ran at 7% to 12% a month, and the last three months ran at 4.0%, 2.6% and 4.9%. A market that stops accelerating is not a market that is falling, and this one has not fallen, yet.
Eight readings, what they say now, and what would count
Crash signals from our own data, as of 2026-09-14
| Reading | Source | Now | Crash signal | Status |
|---|---|---|---|---|
| Index drawdown from high | moonstone Index | 0.0% (high on Sep 14; deepest ever 1.8%) | 20% or more from the high | green |
| Breadth, English cards | Breadth | 37.7% up, 25.5% down over 16 days | 55% or more falling over 30 days | green |
| The leaders | Breadth | top 300 cards: 19% up, 8% down, 73% flat | value-weighted top 300 negative over 60 days | amber |
| Peak timing | The Drawdown Book | 79% of priced cards set their 18-month high since June | a majority more than 20% below their high | amber |
| PSA 10 premium | graded snapshot series | median 4.36x to 4.51x over 2+ weeks on 296 cards; 82 compressed, 99 expanded | median premium ratio below 0.85 | green |
| Graded sales volume | graded snapshot series | PSA 10 sales in the provider window +7.6% (107,429 to 115,548 across 347 cards) | down 20% or more | green |
| Sealed | sealed product tape | 2,007 products: 833 up, 258 down, 916 flat over 30 days; value-weighted +3.9% | value-weighted negative with fallers outnumbering risers | green |
| Registry inflow | Where the New Slabs Go | +2.7% a month, 9,634 slabs a day; Mega era +9.6% | a supply reading; it has no crash threshold on its own | informational |
Six of the eight are green and two are amber. The ambers are the ones worth watching. First, the leaders have stopped moving: over the sixteen days to September 14, 73% of the 300 most valuable English cards were flat while the market beneath them rose 37.7% to 25.5%, and the value-weighted change of that top tier was +1.1% against an index up 4.8%. A rally that narrows to its cheapest cards while its most expensive ones stall is the textbook late-stage shape, and it is also what a healthy market looks like when it pauses; sixteen days cannot distinguish them. Second, The Drawdown Book found that 79% of cards priced $5 and up set their eighteen-month high between June and September. Almost nothing in the market has been tested by a fall, because almost nothing has fallen.
The premium is holding and the slabs are still selling
The readings a bubble would corrupt first are the ones that are not moving. The PSA 10 premium over raw, the multiple that Hard to Gem showed runs from 28x on the hardest cards to 5x on the easiest, has a median of 4.51x across the 296 cards with a real sold tape at both ends of a two-week window, against 4.36x at the start; 82 cards compressed more than 5% and 99 expanded. A crash usually shows up here early, as slab buyers stop paying for the grade before raw buyers stop paying for the card. The number of PSA 10 sales the provider counts in its window rose 7.6% across the same cards. And sealed, which was frozen in August with six gainers in 651 products, now shows 833 of 2,007 products up more than 2% over thirty days against 258 down, a value-weighted +3.9%. Sealed is the layer that fell first and hardest in the last cycle; it is rising.
What is different from the last time
The standing argument for a crash is supply. The Float counted 1.1 million copies on the TCGplayer shelf worth $17.5 million at market, with the newest sets on median shelves of 449 listings per card, and found that a week's change in the shelf predicted nothing about the next week's price. Where the New Slabs Go measured the registry growing 2.7% a month, with the Mega Evolution era at 9.6% and Ascended Heroes at 32.7%, which is the population growth tax accruing in real time on exactly the cards that ran hardest. The Pull-Rate Squeeze and The Print Clock describe a publisher printing more and rationing the chase harder at once. And the breadth study found the newest era already falling, half of Mega Evolution cards down and ME05: Pitch Black at -16.3% value-weighted, while vintage rose four to one. If the market is turning, that is where it is turning first: at the new-product tier, the way every modern set's launch buyers paid the top. The old cards, on thin shelves with closed registries, are not showing it.
Re-read the same eight numbers
The value of a checklist is that it can be re-run. Every reading in the table can be recomputed from the same tables on any date, and the studies behind it, The Drawdown Book, Breadth, The Float and Where the New Slabs Go, print their methods. When two of the eight are red we will say so, in this format, with the date. Until then the honest reading of the tape is: a market at its high, decelerating, broad beneath a stalled top, with its newest product falling and its supply growing, and no drawdown yet to learn from. It does not have a season. It may have a cycle. We have not seen one.
Index figures are the moonstone Pokémon Index daily closes as read on 2026-09-14 (354 points from 2025-03-17; the series is live and a close can be revised by a few hundredths after the fact; a frozen equal-weight basket of fifty blue-chip cards on real TCGplayer histories, homepage series), with the S&P 500 daily close from the same payload; monthly changes compare each month's last close to the prior month's. Drawdown is 1 minus close over running maximum. Breadth and leader readings are from the breadth study (English cards, cheapest printing, 2026-08-29 to 09-14, 60% moves excluded). Peak timing is from The Drawdown Book (2,417 curves, 1,269 priced $5+). PSA 10 premium: cards whose PSA 10 sold median (8+ sales) has a snapshot at two dates at least 14 days apart and a raw market of $5+ at both, premium = median over cheapest-printing raw price at each date; 296 cards; compressed and expanded mean a ratio below 0.95 or above 1.05. Graded sales volume sums the provider's PSA 10 sample size at the first and last snapshot for the same 347 cards. Sealed: the 2,007 tracked sealed products with a price and a 30-day change on 2026-09-15; up and down mean more than 2%; value-weighted compares summed current prices to summed prices 30 days earlier implied by each product's change. Registry inflow is from Where the New Slabs Go. Every reading in the checklist comes from a live series and will differ slightly on any later date; the premium and velocity rows in particular move every day now that the graded snapshot cron refreshes the whole answered set daily. The video that prompted this piece was not transcribed and is not quoted; the question is treated on its own. Not financial advice.