Cardian Research · Market · Valuation · Japanese cards · Chinese cards · One Piece
Rises bring players, crashes decide who stays
On 23 September we wrote: "The market doesn't bounce. It grows in layers." On every index we have, this year's lows sit above the 2024–25 lows, while Japanese raw, Chinese cards and most trainer chase cards are still well below their old highs.
Our Japanese raw index, which follows the 100 most-traded ungraded Japanese Pokémon cards, is 7% above where it started in January 2023. In between it more than doubled, lost 81% over a year, more than tripled and then fell 42% this year. Yet its low this October was twice as high as its low in June 2024.
Our 23 September carousel was built around that pattern: "The market doesn't bounce. It grows in layers." It added "Rises attract people. Crashes filter them" and "the next round starts on a more mature base". The carousel had no numbers in it, so this piece checks it against our longest price histories. Those are English Pokémon since 2021, Japanese since 2023, and One Piece (JP) and Chinese Pokémon, each tracked raw and in PSA 10 (graded a perfect 10 by PSA).
What we said
Card production tripled after 2020 and has stayed high
The carousel started with people, listing "A card that sells for over a million", "An anniversary" and "Even a single viral video" as the sparks that bring newcomers in. Of those newcomers it said most "are here to make money. They leave the moment the market turns", while a few "stay, and start studying the artists and the stories behind cards and become seasoned". We can't count collectors directly, so the closest public number is how many cards The Pokémon Company prints.
Pokémon cards produced each fiscal year, billions (FY20 = April 2019 to March 2020)
Output tripled in the 2021 boom and stayed there through the 2022–24 price slump.
Source: The Pokémon Company production totals as reported by PokeGuardian (27 May 2026); yearly figures are differences between year-end totals
In the year to March 2020 it produced about 3.2 billion cards. Two years later, after the lockdown boom, the figure was 9.1 billion, and it has stayed between 9 and 12 billion every year since, including the two years when card prices fell hardest. Production isn't the same as people, but a company doesn't keep printing three times as many cards for buyers who have left. My guess, and it is only a guess, is that most of the 2021 crowd stayed.
The same carousel said "the next wave of money meets a market with more people, deeper understanding, and less supply in circulation". New supply since the boom has run at around 10 billion cards a year, which matters for which cards hold their price later on.
What the data says
The floor rose in every market we track
Start with English cards, our longest history. The index begins in spring 2021, close to the top of the pandemic boom, so it shows the slow decline after the boom rather than the boom itself.
English Pokémon, rebased so spring 2021 = 100, monthly (October = 12 Oct)
Raw took 30 months to climb back to its 2021 level; PSA 10s took 18.
Source: Cardian Pokémon (EN) indices: the 100 most-traded English cards, from sold prices in the US, raw and PSA 10; monthly before Aug 2026 (cardian.sg/market/method)
Raw cards fell for three years and hit bottom in March 2024, a third lower. PSA 10s bottomed in August 2024, also a third down. Neither came close to zero, and both are now above where they began: raw by 7% and PSA 10s by 80%. That is one full cycle, and Japan gives us two.
Japanese Pokémon, rebased so each index's June 2024 low = 100, daily. End labels: today vs that low
This floor is about twice the last one, but raw is still 60% under its 2023 top: a higher low doesn't promise a higher high.
Source: Cardian Pokémon (JP) indices: the 100 most-traded Japanese cards, from sold prices in Japan, raw and PSA 10 (cardian.sg/market/method)
If you set each Japanese index so its June 2024 low equals 100, this October's lows came in at about 200 for raw and 244 for PSA 10s. The 2026 sell-off took roughly 40% off both, which we covered in our piece on the sell-off. The point here is where it stopped: about twice as high as the last floor.
The same test works for every market. For each index I took the lowest point since its 2026 high and compared it with the lowest point it reached in 2024–25.
Lowest level since each index's 2026 high, vs its lowest level in 2024–25
Chinese cards cleared it by the least, and their fall isn't over, so their low could still drop below it.
Source: Cardian 100-card indices, raw and PSA 10: English Pokémon (sold prices in the US, monthly before Aug 2026), Pokémon (JP) and One Piece (JP) (sold prices in Japan), Chinese Pokémon (sold prices in China) (cardian.sg/market/method)
All seven indices that fell this year stopped above their old floor, from 23% higher for Chinese raw cards to 210% higher for One Piece raw. English raw cards haven't fallen this year at all; they're at a record. One thing to keep in mind: the Chinese indices are still falling, so their low may not be final yet.
What the data says
But three of eight indices never got back to their old high
A higher floor isn't the same thing as a recovery, so the next question is how deep the crashes were and whether prices ever got back. This is the biggest fall in each index's history.
The biggest fall in each index's history: peak to low, then months from the low back to the old high
| Index | Fall | Months down | Months to recover | Now vs top |
|---|---|---|---|---|
| English raw | -33% | 35 | 30 | record |
| English PSA 10 | -34% | 35 | 18 | -8% |
| Japanese raw | -81% | 13 | not yet | -60% |
| Japanese PSA 10 | -61% | 9 | 18 | -22% |
| One Piece raw | -31% | 4 | 17 | -32% |
| One Piece PSA 10 | -48% | 15 | 6 | -44% |
| Chinese raw | -78% | 14 | not yet | -72% |
| Chinese PSA 10 | -76% | 16 | not yet | -64% |
Our English history starts in spring 2021 and our Chinese PSA 10 index starts at its own peak, so those falls began earlier than we can see.
Source: Cardian 100-card indices, raw and PSA 10: English Pokémon (sold prices in the US, monthly before Aug 2026), Pokémon (JP) and One Piece (JP) (sold prices in Japan), Chinese Pokémon (sold prices in China) (cardian.sg/market/method)
The falls were deep and slow, between 31% and 81%, and they took 4 to 35 months to hit bottom. Five indices got back to their old high, 6 to 30 months after the low. Three haven't. Japanese raw cards are still 60% below June 2023, and Chinese cards are 64–72% under their 2023 highs more than three years on.
The carousel's formula ended with "Never back to zero", and an earlier slide read "The market comes back. Not every card does." In our data no index has lost more than 81%, every one is well above its worst level, and the three still short of their highs all peaked in 2023. Whether a price comes back seems to depend on whether it was reached in a sudden spike. The Japanese raw top came after the index rose 139% between 1 May and 3 June 2023, and a price that rose that fast had no buyers underneath to hold it up.
Which cards stay
Crashes filter cards, and the filter isn't random
The carousel also sorted cards into two columns: "COMES BACK" for "Era-defining, irreplaceable cards" and "MAY NEVER RETURN" for "Catch-up pumps, obscure picks, short-term hype". Our release data shows how launch-week chase cards have done since. For each Japanese set from September 2022 to September 2025, we took the five most expensive cards in its first week and compared them with the latest price we have, up to two years later.
Median fall from launch week to the latest price we have (up to 2 years on): the 5 dearest cards in each Japanese set's first week, sets from Sep 2022 to Sep 2025
29% of the Pokémon cards are above their launch price; 6% of the trainers are, and none from 2023 sets.
Source: Cardian release curves, Pokémon (JP): sold prices in Japan, raw, launch week vs latest; 31 sets, 155 cards (item cards not shown). Grey bars: sets released in 2023 only
The difference is clear. Pokémon cards fell a median 32%, and 29% of them now sell above their launch price. Trainer cards, the full-art character cards like Erika's Invitation, Iono or Miriam that were the hottest pulls of 2023, fell a median 67%, and only 3 of 54 are above launch. From the 2023 sets, the boom-year releases, not a single one is. The winners were mostly the obvious icons: Greninja ex from Crimson Haze is up about six-fold from launch week, while Bubble Mew from Shiny Treasure ex and the Pikachu ex from Super Electric Breaker are each up more than 150%.
A card is never priced with more excitement than in its launch week, so this test is tough on every card. My own view, and it's only opinion, is that a trainer card's demand often rests on one wave of popularity. Pikachu, Mew and Charizard have thirty years of collectors behind them, since the game launched in Japan on 20 October 1996. The same pattern shows up today in English cards, where 89% of the Mega-era cards we track fell over the past 30 days against 32% of the WOTC-era ones.
What it means
What to hold through a downturn
Taken together, every market has made higher lows, so selling everything in a crash has usually meant selling close to the bottom. But the floor belongs to the market, not to your card. Our indices are rebuilt every month from the 100 most-traded cards. A card that stops trading simply drops out, so it no longer drags the index down.
So the useful question in a downturn isn't whether the market will come back, but whether your card is one the next crowd will still want. In my view that comes down to three questions. Would people want it if its price had never moved? Is it the Pokémon or the artwork they chase, rather than a trend? And did you buy it in a month when everything rose at once? If the card passes, our data says holding on has been the better choice. If it doesn't, a crash is a good time to swap it for something older or more iconic.
We sold our Japanese PSA slabs during the run-up and have been restocking since the fall, because, as we wrote on 23 September, "We buy when people are scared and sit on our hands when everyone's a genius". After running these numbers I'd add one rule to that: buy the cards that stay, because the market's floor won't hold up a card nobody wants.