Vintage One Piece · Research & Archive
Special report · November 8, 2026

Special report · 1999 Carddass Hyper Battle

The First Stage Oda copyright error

Some cards in the first One Piece card set misprint Eiichiro Oda’s name as instead of . About one in four audited PSA-graded copies carries it, on every design checked, and Bandai’s corrected art was in use by December 18, 1999.

Result · 180 PSA-graded First Stage cards audited

26.1%Error · 47 cards

73.9%Correct · 133 cards

A share of audited graded copies, not of the print run. The range allows for sample size.

Every design checked
Both spellings turn up on all eight designs audited, holo and non-foil.
Corrected art by Dec 18, 1999
A Bandai promo handed out at Jump Festa 2000 already reads .
First Stage only
More than 50 later-stage cards checked. None carries the error.

Jump to: How to spot it · For collectors · All 180 certificates

Key findings

  1. About one in four graded copies carries the misprint. 47 of 180 audited PSA certificates read . It is a share of graded copies, not a print-run figure.

    Supported
  2. It shows up on every design checked. Both spellings appear on all eight designs audited, and no design stands out (p = 0.87). That points to the card back every First Stage card shares.

    Confirmed
  3. The misprint came first. The vending header that advertised First Stage reads . A Bandai promo with the same back, handed out December 18 to 19, 1999, reads , as does every later-stage card checked.

    Supported
  4. Error copies turn up together. Cards graded in the same PSA submission tend to share a spelling, across designs. That fits cards bought from the same vending stock.

    Supported
  5. Error copies may reach PSA 10 less often. 1 of 47 error copies is a PSA 10, against 16 of 133 correct copies. The gap could still be chance (p ≈ 0.07).

    Hint
  6. Two questions remain open. When First Stage card printing switched to the corrected art, and whether every error card comes from one early run.

    Open

Confirmed seen directly · Supported the evidence points this way · Hint suggestive, could be chance · Open not yet known

01 · How to spot it

Read the last character of Oda’s name on the back

Every First Stage card shares one back, with the copyright line under the ONE PIECE banner. On error copies the fifth character, just before the slash, has on its right side instead of .

Figure 1Two C01 Luffy cards, consecutive certificates, different spellings

Scan of the copyright line on PSA certificate 162195362, reading ©尾田栄一朗/集英社. The last character of the name is 朗.
ErrorPSA cert 162195362
Scan of the copyright line on PSA certificate 162195361, reading ©尾田栄一郎/集英社, the correct spelling.
CorrectPSA cert 162195361
Printed on error copies
Read rō. Right side is , two short bars inside a tall box.
Oda’s actual name
Also read rō. Right side is , shaped like a 3 on a stem.
The back of a First Stage card in a PSA slab, with the copyright line under the ONE PIECE banner outlined.

Where to look. The copyright line under the ONE PIECE banner, outlined. Every First Stage card shares this back.

PSA certificates 162195362 (PSA 3) and 162195361 (PSA 5): same design, same submission, different spelling. PSA images © Collectors Universe, Inc.
  1. Graded cardOpen the certificate on psacard.com and read the back image. PSA does not list this as a variety, so the label won’t say.
  2. Raw cardAsk the seller for a sharp photo of the line under the banner.
  3. Later setsNo need to check. Every later-stage card checked reads .

02 · Timeline

The misprint came first, and corrected art was out by mid-December 1999

Three dated sources bracket the change: the set’s own vending header, a Bandai promo from December 1999, and the sets that followed.

Figure 2The correction window

Corrected art first appears inside the shaded window, between First Stage’s November launch and Jump Festa 2000. It dates the art change, not when First Stage card printing switched. Exact release days are not published, so the window opens at the start of November.

Nov 1999 · First Stage launch

Vending header reads

The artwork that advertised the set, with the same copyright line as the card back. Two copies checked, one graded by CGC.

Dec 18–19, 1999 · Jump Festa 2000

Promo C-J1 reads

Bandai’s gift to booth visitors at Tokyo Big Sight, with the same back as First Stage. Every copy checked agrees, raw and graded.

Jan 2000 onward · Second Stage and later

Later stages read

More than 50 cards checked from Second Stage onward. None carries the error.

Not yet known: when First Stage card printing switched to the corrected art, and whether every error card comes from one early run.

The header in detail

Each Carddass machine displayed a printed header advertising the set inside. The First Stage header shares the card back’s copyright element: the same typeface, separators and position under the From TV animation logo.

Figure 3The First Stage vending machine header, two copies

Flat scan of the First Stage Carddass 100 vending machine header: Luffy, Zoro and Nami above a C01 Luffy card, the ONE PIECE First Stage logo and the price panel. The copyright line under the logo is outlined.

Flat scan

A second First Stage header graded CGC EX/NM 6, certificate 6164833003, in its holder. The label reads One Piece First Stage, Carddass Hyper Battle (1999) Japanese, Vending Machine Mount, 100 Yen Version. The copyright line is outlined.

CGC EX/NM 6, cert 6164833003. Photo: mediagiraffes on Instagram

Close-up of the copyright line on the flat-scanned header, reading ©尾田栄一朗/集英社. The last character of the name is 朗.
ErrorFlat scan
Close-up of the copyright line on the CGC-graded header, photographed through the holder, reading ©尾田栄一朗/集英社. Softer, but the last character is 朗.
ErrorCGC 6164833003, through the holder
Format
Carddass 100 machine header: , ,
Sold by
,
Product code
751293-0074328-100
JAN
4902425743288
Imprint
BANDAI 1999 MADE IN JAPAN
Copyright
Two different copies, both reading . Select either one to open it full size.

One copy is a sharp flat scan. The other is graded CGC EX/NM 6 (cert 6164833003) and reads through the holder. Its product code and JAN match a separate copy listed on Yahoo! Auctions, and the JAN carries Bandai’s 4902425 prefix.

About First Stage Set facts, release date and how long it ran

First Stage, subtitled VERSUS!! Buggy Pirates (), opened Bandai’s Carddass Hyper Battle line. Grand Line Wiki lists it as the first official One Piece trading card set. It reached vending machines in November 1999, a few weeks after the TV anime premiered on Fuji TV on October 20, 1999, and its back carries the anime’s “From TV animation” logo.

Released
November 1999. No exact day has been published.
Sold as
Carddass 100 vending machines, ¥100 for 4 cards
Cards
42 designs: 34 character cards (C01–C34) and 8 special cards (S01–S08)
Holos
6 prism designs: C01 Luffy, C07 Zoro, C14 Shanks, C24 Buggy, S01 Luffy Pirates, S05 Nami. The other 36 are normal cards.
Story
Romance Dawn and Orange Town arcs
Replaced by
Second Stage, January 2000

A two-month window

Bars run from a set’s release month to the next vending set. Release months only; exact days aren’t published. Grand Box was a boxed deck, not a vending set.

Second Stage arrived in January 2000, so First Stage was the newest set in machines for about two months. Later stages followed every two to three months. Unsold First Stage cards could have stayed in machines longer, and no print-run figures are public.

03 · The numbers

One in four audited copies, spread evenly across the set

We read the copyright line on PSA’s own back scans of 180 graded First Stage cards, covering all six holo designs and two non-foil rookies. 47 carry the error. Allowing for sample size, the share among graded copies most likely sits between 20% and 33%: about three correct copies for every error copy.

Figure 4Every certificate in the audit, by design

Error Correct

Each square is one PSA-graded card. Within each row, error copies come first, then by grade. Click any square to open its PSA record.Tap any square for its details. The full list, with PSA links, is in the appendix.

No design stands out

Design rates run from 12.5% to 33.3%, all within chance of the pooled rate (chi-square, p = 0.87). The six holos come in at 27.1% (42 of 155) and the two non-foil designs at 20.0% (5 of 25). An even spread is what a change to the shared back would produce.

Figure 5Share of error copies by design, with 95% intervals

Design rateAll 180 cards, 95% interval 20.2–33.0%
Dots show each design’s share of error copies; lines show the 95% interval. Every interval overlaps the pooled band. Small samples give wide intervals (C06 Zoro has 9 certificates).

Fewer PSA 10s among error copies, possibly by chance

Error copies average PSA 7.45, against 7.92 for correct copies. One of 47 error copies is a 10, against 16 of 133: 2% against 12%. A gap this size would appear by chance about 7% of the time, so treat it as a hint.

Figure 6Grade distribution, as a share of each version

Error copies (47)Correct copies (133)
Show the counts
Bars show the percentage of each version at each PSA grade, so the two sets compare fairly despite different sizes. No certificates in the sample graded PSA 2.

Error copies arrive together

PSA numbers certificates in order, so cards graded together sit close in number. Grouping certificates within 10 numbers of each other gives 28 groups of two or more cards. Six are all errors, against 1.3 expected by chance (p ≈ 0.0007). Inside groups, 10 of 34 pairs of different designs are both errors, against 2.4 expected (p ≈ 0.0002).

GroupCertificateCardPSA
1109561478C14 Shanks8
109561480S05 Nami7
2149687222C07 Zoro7
149687224C24 Buggy8
3167756311C01 Luffy6
167756312C01 Luffy5
4169949186C01 Luffy9
169949187C07 Zoro8
169949188S05 Nami9
5173947425C06 Zoro9
173947427C14 Shanks9
173947428C24 Buggy9
6176050842C24 Buggy8
176050844S05 Nami7

The six all-error groups. Five of them mix designs.

That fits cards bought from the same vending stock. A refill box carries the whole set, so a box from an early lot would be all across designs; a sheet-position fault would mix both spellings in every box. Consecutive numbers show cards were graded together, not printed together, so the vending-stock link is an inference.

04 · Why it happened

Best explanation: a typo in the original art, corrected during the run

and are both read rō and sit in the same row of the JIS character set (47-15 and 47-26), close together in typing conversion lists. is also a common way to write Ichirō; Ichiro Suzuki’s legal name is . It is a plausible mechanism, not a documented one.

#ExplanationVerdict
1Typo in the original art, corrected during the runHeader reads ; the December 1999 promo and later stages read ; the rate is even across designs.Best supported
2Two production streams, only one correctedA reprint lot or second printer kept the old art. Fits the clustering equally well; dated or regional evidence would separate it from the first.Possible
3The error came with licensor-supplied logo artThe copyright line travels with the anime logo. If so, other licensees’ late-1999 products would show it too.Untested
4A sheet-position or plate faultThe header carries the error, and errors cluster across designs. A fault in one part of a sheet would do neither.Not supported
5 came first and laterThe header reads .Ruled out

05 · For collectors

What it means for collectors

  • Expect it on any First Stage design. Every card shares this back. Only 8 of 42 designs have been audited, and all eight have error copies.
  • Ask about the rest of the lot. Error copies cluster. If a seller’s card reads , their other First Stage cards may too.
  • PSA 10 error copies are rare so far. One among 47 error copies in this sample, against 16 among 133 correct copies. The sample is small.
  • Headers carry it too. Both First Stage vending headers checked read .

Help settle what’s open

  • A First Stage header that reads . It would point to a reprinted, later restock.
  • A sealed box or a documented single-box lot. Under the leading explanation it should be one spelling throughout.
  • Other licensees’ One Piece products from late 1999. If they read , the error came with the licensor’s art.

Have one of these, or an error copy of your own? Send the cert number or photos to @vintageonepiececards. Designs not yet covered help most.

06 · Method

Method, limits and sources

How we did it

  • Collected 180 PSA certificates for eight of the set’s 42 designs: the six holos (C01 Luffy, C07 Zoro, S05 Nami, C14 Shanks, S01 Straw Hat Crew, C24 Buggy) and two non-foil rookies (C06 Zoro, C09 Nami). Each was matched to PSA’s record for card and grade.
  • Read the copyright line on every card from PSA’s own back scan. A second, blind read of the first 179 found three error copies the first pass missed (below), each confirmed on PSA’s full-resolution scan.
  • One certificate was reported by a collector after the error went public. Its back reads with the slab number visible, but it was not collected blind, so the appendix flags it “added later”. Without it the rate is 25.7%. It is also the sample’s only PSA 10 error copy.
  • Intervals are 95% Wilson intervals. The design comparison is a chi-square test checked against 20,000 shuffles. Grade tests are shuffle tests run within each design.
  • Group test: certificates sorted by number join a group when within 10 numbers of the previous one. All-error groups and cross-design error pairs were compared with 20,000 within-design shuffles (fixed seed), leaving out the added-later certificate.
  • Header and promo: we read the copyright line on two First Stage headers (a flat scan, and a CGC copy through its holder), checked the header’s JAN check digit, and read the same line on several C-J1 promos, one in hand and the others, raw and graded, from photos.
CertificateCardGradeFirst passCorrected
175777684S05 Nami8CorrectError
175894645S05 Nami8CorrectError
173947427C14 Shanks9CorrectError

Limits

  • A sample of accessible PSA certificates with back scans, not a census or a print-run estimate. C01 Luffy makes up 64 of the 180.
  • The first 179 cards were graded before the error was known, so submitters were not picking for it. Raw cards may differ.
  • Eight of 42 designs are audited. The later-stage check covers more than 50 cards, not every design.
  • PSA has not designated the error as a variety. The classification is ours.
  • The header evidence rests on two copies. The header is dated only BANDAI 1999; treating it as launch art follows from its role as the set’s point-of-sale artwork.
  • The group counts were chosen after seeing the data. Treat them as a strong indication, not proof.
  • No Bandai notice or recall was found, and the 1999 print contractor is not documented.

Sources

Version history
  • v2.0, November 8, 2026. Final report. Rewritten and restructured: a results summary, key findings with confidence labels, a timeline-first dating section, and one place for method, limits and sources. Statements about correction timing, production runs, the print-run share and PSA 10 scarcity now match the evidence. Certificate “batches” are now “groups”. Added section navigation. More C-J1 promos checked, raw and graded, and every one reads . No certificate data or calculations changed.
  • v1.2, November 7, 2026. Added the First Stage vending header (two copies, one CGC-graded) and the C-J1 Jump Festa 2000 promo. Together they put the error in the launch artwork and the fix before mid-December 1999. Re-ran the batch test with counts that look at error copies only, replacing the weak v1.1 result. Added the correction-window timeline, ranked explanations, the character-set mechanism, open questions, new limits and sources. Updated the summary, evidence table and collector notes. The page now leads with the dating evidence, and the card analysis follows it. No certificate data changed.
  • v1.1, November 7, 2026. Added one collector-reported certificate, flagged “added later”.
Appendix: all 180 certificates

Back to key findings

Card · certGradeSpellingEvidence