Cardian Research · Market · Buyer's guide · Japanese cards · One Piece · Chinese cards
Wobble or crash? Three tests, checked against every fall since 2023
A crash needs three things to go wrong at the same time: breadth, depth and buyers. Here I define each one using our own indices, test it on past falls in four markets, and use it to score 2026.
Between mid-2023 and mid-2024, four of the indices we track lost between 61% and 81% and stayed down for most of a year. Since then there have been plenty of frightening falls, including about 40% in Japanese Pokémon this year and 51% in One Piece slabs, and none has looked like 2023–24 on the numbers that matter.
Every time prices drop, the same question comes up: is the market crashing? On 5 September we wrote on Instagram: "Cards have no official definition — so here's mine. A crash needs all three", and the three were breadth, depth and liquidity. This piece turns each of those into a test with a number, checks each cut-off against past falls, and scores each market today.
The indices are Cardian's own: for each market, the 100 most-traded cards priced from real sales, once for raw cards (ungraded) and once for PSA 10s (graded a perfect 10 and sealed in a plastic slab). I use Japanese Pokémon, One Piece (Japanese), English Pokémon and Chinese Pokémon.
What happened
Depth: crashes go 50% down and stay there
The first test is depth. It asks whether an index is at least 50% below its highest level of the past 12 months, and whether it has stayed there for 90 days. I measure from the 12-month high rather than the all-time top because you can check it on any price chart, and an old bubble shouldn't count against a market forever. Japanese raw cards, for example, are still 60% below their June 2023 peak, but that tells you about 2023, not about today.
Deepest fall below the index's 12-month high, and days spent 50% or more below it
The four crashes spent 216 to 415 days at least 50% down. The only other index to get there, One Piece PSA 10s, stayed 10 days.
Source: Cardian indices: for each market the 100 most-traded cards, from sold prices, raw and PSA 10: Pokémon (JP), One Piece (JP), Pokémon (CN) and Pokémon (EN) (cardian.sg/market/method)
The 50% line sits in a very wide gap between crashes and other falls. Japanese raw cards spent 415 days at least 50% below their 12-month high, Chinese raw 378, Chinese PSA 10s 367 and Japanese PSA 10s 216. In every other fall we've recorded, only one index crossed the line at all: One Piece PSA 10s, for ten days in late September, after the index had more than tripled between March and June. A 51% drop after a run like that hurts, but it was a spike falling back, and it lasted days rather than months.
What happened
Breadth: in a crash the selling doesn't stop
Depth alone can be misled by a few expensive cards, so the second test is breadth, meaning how many cards are falling. Each day I take the cards in the index that moved more than 1% over the past week and work out what share of them fell. Averaged over four weeks, the test "trips", or is triggered, when falling cards have been the majority for 180 days in a row.
Cards in the raw index rising vs falling over 7 days (by more than 1%), 4-week average, weekly
In 2023–24 the falling side led for 234 days straight. The bars also shrank as fewer cards traded, which is the one crash where that happened.
Source: Cardian Pokémon (JP) indices: the 100 most-traded Japanese cards, from sold prices in Japan, raw and PSA 10 (cardian.sg/market/method)
Japanese raw shows what that looks like. In 2023–24 falling cards outnumbered rising ones for 125 days, paused for two weeks, then led again for another 234. The bars also shrink as fewer cards changed price, which matters for the third test. The worrying part is at the right-hand end of the chart: falling cards have now led for 143 days, the longest stretch since that crash.
Longest stretch in which falling cards outnumbered rising ones (4-week average), days
No other fall in our daily indices passed 147 days, so six months (180) sits in the gap. English cards passed it in a slow slide that never went deep, which is why one test alone isn't enough.
Source: Cardian indices: for each market the 100 most-traded cards, from sold prices, raw and PSA 10: Pokémon (JP), One Piece (JP), Pokémon (CN) and Pokémon (EN) (cardian.sg/market/method). English indices were priced monthly before Aug 2026
Across every market, falling cards led for between 195 and 366 days in each crash. In the other falls in our daily indices, the longest stretch was 147 days, in One Piece PSA 10s from August to December 2024. The sharpest dip of early 2025, a 25% drop in Japanese PSA 10s, had falling cards in the lead for just 26 days. Six months sits in that gap, although four crashes is a small sample, so the line is less precise than it looks.
The English index shows why one test isn't enough. From 2021 English cards fell by a third over three years, with falling cards in the lead for 275 days on monthly prices, yet the index was never more than 22% below its 12-month high. That's a slow bear market (a long, steady fall) that wears holders down, not a crash. One Piece slabs this year are the opposite: deep but brief. In our whole history, the two tests have only been tripped together between February and August 2024, and only in the four indices that crashed.
What's happening now
Liquidity: the test our data can't score yet
The third part of the September framework was liquidity, which the post described as "Nobody bids. Dealers slash buylists, boxes trade below retail, volume dries up." Here I have to be honest about our data. Our indices record prices, and the data I'm working from only has today's sales counts, with no history, so I can't test a cut-off against 2023–24.
I tried the obvious substitute: counting how many of the 100 cards move by more than 1% in a week, since a card nobody trades doesn't move. It doesn't work.
Index cards whose price moved more than 1% in a week, average over the 4 weeks before the top and before the low
If this worked as a liquidity test, every crash would show a big drop. Only Japanese raw did, so I don't use it.
Source: Cardian indices: for each market the 100 most-traded cards, from sold prices, raw and PSA 10: Pokémon (JP), One Piece (JP), Pokémon (CN) and Pokémon (EN) (cardian.sg/market/method)
The count collapsed in the Japanese raw crash, from 85 cards a week to 35, but barely moved in the other three. It also fell just as hard in two One Piece raw dips of about a third that never came close to a crash.
So for liquidity I use readings I can't yet test, and treat them as background rather than a score. Japanese PSA 10 sales over the past month rank in the bottom third of the past year, even as prices have recovered, while raw sales are about normal. All 21 Japanese One Piece booster boxes in our sealed tracker still sell above retail, the cheapest at 2.3 times.
What we said
After 5 September, Japanese PSA 10s fell another 22%
On 5 September the first slide of our carousel read: "My read: a wobble, not a crash — yet." Its index page added: "Index verdict: correction, not crash. Line in the sand: ¥110k (Jan 2026 breakout). Below that, I change the call." That ¥110k was on pokeca-chart's 100-card index, about 31% below its April peak of ¥158.7k, and on our own index PSA 10s were then 19% below their April top.
Japanese Pokémon, % below each index's 12-month high (both set in April), daily
PSA 10s passed −31% on 19 Sep and bottomed at −37% on 1 Oct, 13 points short of the depth test.
Source: Cardian Pokémon (JP) indices: the 100 most-traded Japanese cards, from sold prices in Japan, raw and PSA 10 (cardian.sg/market/method)
On our index, PSA 10s passed the same 31% fall on 19 September and kept going to 37% below the top on 1 October, while raw cards fell another 21% after 5 September. About the selling everyone expected around the 30th anniversary, the same post had said "The dip you're expecting gets front-run — and shallow". Our next carousel, on 25 September, opened with "The Pokémon card crash has a cause, an end date and a buy signal". The index turned up on 1 October, which is the story told in our piece on the sell-off.
On this piece's tests, PSA 10s bottomed 37% below their 12-month high, 13 points short of the depth line, and falling cards led for 77 days at most, compared with the 180 that breadth needs.
What's happening now
No market trips a test today, but one is getting close
Scorecard on 13 Oct 2026 (English and Chinese to 12 Oct)
| Market | Off 12-mo high, now / low | Falling-led, days | Sales vs usual |
|---|---|---|---|
| Japan PSA 10 | 22 / 37% | 65 | thin |
| Japan raw | 39 / 42% | 143 | normal |
| One Piece PSA 10 | 44 / 51% | 105 | normal |
| One Piece raw | 32 / 33% | 67 | thin |
| China PSA 10 | 27 / 27% | 40 | no count |
| China raw | 15 / 15% | 101 | no count |
| English PSA 10 | 8 / 10% | 0 | no count |
| English raw | 0, record high | 27 | no count |
None is 50% down for 90 days or has a 180-day falling-led run. If Japanese raw's run continues, it reaches 180 on 19 Nov. Sales: last 30 days against the 90-day average, ranked over the past year.
Source: Cardian indices: for each market the 100 most-traded cards, from sold prices, raw and PSA 10: Pokémon (JP), One Piece (JP), Pokémon (CN) and Pokémon (EN) (cardian.sg/market/method); sales counts for Japanese and One Piece only
Japanese raw is the one to watch. Falling cards have led since 24 May, and if that continues, the breadth test trips on 19 November. Depth is further off: the index would have to fall another 18% from here to reach half its April top. One Piece PSA 10s went deepest this year, but their falling streak is 105 days. Chinese raw has had falling cards in the lead for 101 days, yet it's only 15% below its high.
The markets are also moving differently from 2023–24, when every Pokémon market we track daily fell together. This year Japanese cards and One Piece fell while English raw cards set a record on 12 October, so whatever this is, it isn't the whole hobby breaking at once.
What's next
Your three-question checklist
- Depth: Is it at least 50% below its highest price of the past 12 months, and has it stayed there for three months? Crashes spent 216 to 415 days there, and no other fall lasted more than ten.
- Breadth: Have more cards been falling than rising, over four-week stretches, for six months straight? Crashes did it for 195 days or more, and other falls in our daily data lasted 147 at most.
- Liquidity: Are buyers leaving, with this year's boxes below retail, shops cutting buy prices on old and new cards at once, and sales thinning month after month? Watch this one, but don't score it until we have history to test it against.
Three yeses is a crash. One is a wobble, however much it hurts, and two is where I'd plan for a long, slow market.
I'll repost this scorecard whenever a test changes, with the thresholds left exactly where they are today so I can't change the rules afterwards. I'd still give the advice from 5 September: "Buy only what you'd be happy holding through a further 50% drop", and "Don't borrow to buy cards. Ever."