Pokemon Card Japan Sold Prices — Raw vs Graded per Card

What it does: Finds sales of a card name on Yahoo! Auctions and Mercari, works out which exact card each sale was, and returns sold prices per card, raw and graded apart.

You enter: card names in Japanese, e.g. リザードン (search broadly; Pokemon by default, other games selectable)

You get: per card: set code, number, rarity, median sold price for raw and for PSA/BGS graded copies with the sales behind each, graded-to-raw multiple, Yuyu-tei shop price when in stock; plus how many sampled sales matched a card. Up to 60 cards per name.

Price: $0.02 per card name. +$0.002 per listing if you also want the list. Charged only when real sales resolved to at least one card.

Example: enter リザードン → 26 cards identified from 200 sampled sales (100 per site) · メガリザードンXex (m02a 223/193): raw median ¥4,800 · graded ¥15,800 · ×3.29 · shop asks ¥7,980

Unofficial — not affiliated with Yahoo! Auctions, Mercari or Yuyu-tei. Reads public pages only.

What you get

cards[]	One entry per identified card: setCode, cardNumber, rarity, name, and separate raw and graded sold statistics (min / p25 / median / p75 / max / average, each with its own n)
gradedPremium	graded median ÷ raw median per card, with both sample sizes. Measured range on real data: ×1.6 to ×5.03
shopAsk	What a Japanese retailer (Yuyu-tei 遊々亭) asks for the same card today, and the gap to the realized median — only when the row is in stock (see below)
keywordBand	The price band of the whole search term, raw vs graded, over every comparable listing — the denominator
coverage	matchedCount / unmatchedCount / coverageRate, overall and per marketplace
sources	Per-source availability, hit count, sample size and source URL
catalogStatus	ok, no_match (the shop has no card by that word — with a catalogNote saying so) or unavailable (the catalog could not be read)

USD conversion included. No login, no API key, nothing stored between runs. The default Overview shows the per-card results and sample coverage; Individual sales shows the optional listing rows. The JSON export keeps every field, including query, game and checkedAt on both row types. Both table views also retain the search name, game code and UTC check time. Summary table exports name them summaryObservation.query, summaryObservation.game and summaryObservation.checkedAt. Individual sales use listingObservation.query, listingObservation.game and listingObservation.checkedAt. They repeat the matching top-level values.

Three bases, tagged on every single statistic

Every block in the output states all three. Ignoring them is how a price product becomes a lie:

Tag	Values	Why it matters
priceBasis	sold / asking	Yahoo! and Mercari are realized sales; the shop catalog is an asking price
gradingBasis	raw / graded	Never pooled. sv10 114/098 measured raw ¥2,999 vs graded ¥15,100 = ×5.03. 32–49 % of listings are graded, so one blended median describes no card that exists
identityBasis	card / keyword	A keyword band for リザードン spans forty different cards. It is a useful denominator, never a card price

Out-of-stock shop prices are shown but never differenced. An out-of-stock 提示価格 keeps its last value and drifts far above the market — xy08bs 062/059 asked ¥178,000 against a ¥17,750 sold median (×10.03). Publishing that gap would invent an arbitrage that does not exist, so shopAsk.usable goes false and no difference is computed. (In the live run below, three cards hit this — including one asking ¥128,000 with no stock.)

Coverage is published, not hidden

Marketplace titles are free text, so not every listing can be resolved to a card. Measured over 360 real listings:

	Resolved to exactly one card
Yahoo! Auctions	80.4 % of comparable listings
Mercari	48.9 % (shorter titles, usually no collector number)
Combined	62.6 % of comparable listings — 55.3 % of all listings

Identification precision, hand-audited: 99.0 %.

Unresolved listings are not discarded — they fall back to keywordBand with their own n, and coverage.coverageRate always tells you how much of the sample made it. The only listings dropped from every statistic are lots (まとめ / オリパ / N枚 / …) and non-Japanese printings, counted as excludedListings: a 30-card lot at ¥30,000 is not a sold price for one card.

Pricing — $0.02 per card name

Event	Price	When
card-analyzed	$0.02	Once per card name analyzed
listing-scraped	$0.002	Only if you enable Include individual listings

A default run (one card name) costs $0.02 and takes about 5 seconds. A card name is charged only when real sales were found and at least one of them resolved to a catalog card. A word the shop catalog does not know (an English card name, for example) returns the keyword band with catalogStatus: "no_match" and is not charged, because a band that spans forty different cards is not the per-card price this Actor sells.

You pay per card name; there is no monthly fee and no minimum.

Input

Field	Example	Notes
cardNames	["リザードン"]	Japanese. Search broadly — see below. Max 10 per run, $0.02 each
game	"poc"	poc Pokemon (measured default) / ygo / opc / dm / ws / ua / vg / digi / bs / gcg
maxSoldPerSource	100	Sold listings sampled per marketplace (30–300)
minSoldPerCard	1	Hide thin cards from cards[]. Never changes a denominator — coverage and keywordBand are unfiltered
includeIndividualListings	false	Every listing with the card it resolved to and why (+$0.002 each)
convertToUsd	true	Adds USD next to JPY

Card names are Japanese

The shop catalog is a Japanese one, so an English name finds nothing in it: measured 2026-08-18, charizard returned 0 catalog cards while リザードン returned 126. Marketplaces still return listings for the English word, but none of them can be resolved to a card — the record comes back with catalogStatus: "no_match" and is not charged.

Measured on the catalog, 2026-08-18 (Pokemon, game: "poc"):

English	Enter this	Catalog cards
Charizard	リザードン	126
Pikachu	ピカチュウ	312
Mewtwo	ミュウツー	128
Eevee	イーブイ	108
Rayquaza	レックウザ	73
Umbreon	ブラッキー	38

For any other card, the Japanese name is the one printed on the Japanese release — Bulbapedia lists it on every card and Pokemon page ("Japanese name"), and the shop's own site (yuyu-tei.jp) is searchable with it.

Search broadly, not narrowly. Use リザードン, not リザードンex. The shop catalog is looked up with the same word, and a narrow word leaves prints like メガリザードンXex out of it — 11 sales were lost that way when measured. Individual cards are separated afterwards, locally, at no extra request cost.

Run the same watchlist again

Save the input as an Apify Task, then rerun it or add a daily or weekly schedule to match your pricing work. Keep cardNames, game, maxSoldPerSource and minSoldPerCard unchanged when comparing observations. For example, a shop can save ["リザードン", "ピカチュウ"] with game: "poc", check the same cards each week, and keep the JPY medians and sample sizes in its own price history. This Actor does not store history or return only new sales; every run reads another bounded sample.

1. After a run succeeds, export its default dataset as JSON. Retain the run ID and the saved input alongside it. A failed run can have partial rows; do not treat them as a complete watchlist refresh. If fewer summary rows than requested names arrive, check the run log: the time budget can skip later names without charging them.
2. Select type: "card_sold_comps". checkedAt is the UTC time when that query's observation was assembled after the source requests. It is not the time a card sold, and different names can have different times. Before comparing prices, inspect catalogStatus, each source's available / error, coverage.coverageRate, and both raw and graded n. A partial-source result is a different sample from a full-source one.
3. For each entry in cards, compare the same game + cardKey + gradingBasis over time, with priceBasis: "sold". Use query and the saved input to distinguish overlapping searches and different sample settings. Keep snapshots under checkedAt (and run ID for retries). cardKey is setCode|cardNumber|rarity within that game's current Yuyu-tei catalog; it is not a global card ID. A catalog correction can change it. Graded medians pool the observed grades; check grades before attributing a price change to the market rather than a different mix.
4. To maintain an auditable sales collection, enable includeIndividualListings (+$0.002 per listing). Select type: "listing" and deduplicate sales with source + id. Different marketplaces can reuse an ID, and overlapping names or later runs can return the same sale. For a history of observations, keep game + query + source + id + checkedAt instead. Listing rows carry their summary's game and checkedAt; cardKey, matchVerdict and the two countedIn* flags explain whether the sale contributed to a card price, only the keyword band, or neither. Deduplicating your export does not change the per-run charges.

A missing card is not a deletion or a zero price. It may fall below minSoldPerCard, leave the bounded sample, fail to match, or fall beyond the 60-entry cards limit. cardsReported counts cards passing the threshold before that limit; cardsTruncated: true says the array was capped. coverage.identifiedCards counts identities before the threshold, and coverageRate uses comparable listings after lot / foreign-print exclusions. Do not replace an earlier card price with zero when it disappears. Likewise, shopAsk.usable: false means the displayed asking price is unusable for the comparison; an out-of-stock row is not evidence of a new sale.

Compare JPY values for a price history. Optional USD values also move with the exchange rate, and these observations do not promise a fixed sale-date window or complete marketplace history.

Output example

Measured on 2026-08-07 (real run) — cardNames: ["リザードン"], maxSoldPerSource: 100, minSoldPerCard: 2. 200 sold listings sampled, 38 lots excluded, 162 comparable, 80 resolved to 26 distinct cards.

{
  "type": "card_sold_comps",
  "query": "リザードン",
  "game": "poc",
  "coverage": {
    "soldListings": 200,
    "excludedListings": 38,
    "comparableListings": 162,
    "matchedCount": 80,
    "unmatchedCount": 82,
    "coverageRate": 0.4938,
    "identifiedCards": 26,
    "byVerdict": {"NO_HANDLE": 66, "MATCH_NUM": 49, "REJECT_LOT": 38,
                  "MATCH_NAME": 31, "AMBIGUOUS_NAME": 10, "NO_MATCH": 5,
                  "AMBIGUOUS_NUM": 1},
    "bySource": {
      "yahoo_auction": {"soldListings": 100, "comparableListings": 79,
                        "matchedCount": 49, "coverageRate": 0.6203},
      "mercari":       {"soldListings": 100, "comparableListings": 83,
                        "matchedCount": 31, "coverageRate": 0.3735}
    }
  },
  "keywordBand": {
    "identityBasis": "keyword",
    "raw":    {"gradingBasis": "raw", "priceBasis": "sold", "n": 108,
               "soldJpy": {"min": 420, "p25": 2075, "median": 4800,
                           "p75": 14500, "max": 320000, "average": 23260}},
    "graded": {"gradingBasis": "graded", "priceBasis": "sold", "n": 54,
               "soldJpy": {"min": 5780, "p25": 15050, "median": 27800,
                           "p75": 84150, "max": 900000, "average": 78529},
               "grades": {"PSA10": 49, "PSA9": 3, "ARS10": 1, "ARS9": 1}},
    "gradedShare": 0.3333
  },
  "cardsReported": 18,
  "cards": [
    {
      "cardKey": "m02a|223/193|MA",
      "setCode": "m02a", "cardNumber": "223/193", "rarity": "MA",
      "name": "メガリザードンXex",
      "soldListings": 10,
      "raw":    {"identityBasis": "card", "gradingBasis": "raw",
                 "priceBasis": "sold", "n": 4,
                 "soldJpy": {"min": 4350, "p25": 4688, "median": 4800,
                             "p75": 4825, "max": 4900, "average": 4712},
                 "bySource": {"yahoo_auction": 4, "mercari": 0}},
      "graded": {"identityBasis": "card", "gradingBasis": "graded",
                 "priceBasis": "sold", "n": 6,
                 "soldJpy": {"min": 6600, "p25": 15275, "median": 15800,
                             "p75": 15950, "max": 16200, "average": 14250},
                 "grades": {"PSA10": 5, "PSA9": 1},
                 "bySource": {"yahoo_auction": 5, "mercari": 1}},
      "gradedShare": 0.6,
      "gradedPremium": {"multiple": 3.29, "rawMedianJpy": 4800, "rawN": 4,
                        "gradedMedianJpy": 15800, "gradedN": 6,
                        "basis": "graded_sold_median / raw_sold_median"},
      "shopAsk": {"priceJpy": 7980, "stockStatus": "in_stock",
                  "priceBasis": "asking", "usable": true,
                  "detailUrl": "https://yuyu-tei.jp/sell/poc/card/m02a/10459",
                  "vsRawSoldMedian": {"basis": "asking_vs_sold",
                                      "diffJpy": 3180, "diffPct": 66.2}}
    }
  ]
}

メガリザードンXex is exactly the card a narrow catalog query would have missed. (That run predates the catalogStatus field. This abbreviated historical example also omits checkedAt; current summary rows include both fields.)

Who this is for

• Resellers and arbitrage — the Japanese realized price is the number that decides whether an import is worth it, and gradedPremium tells you whether the margin is in the card or in the slab
• Graders — is submitting this card worth the fee? gradedPremium with both n values answers it per card, not per set
• Collectors — what a card really goes for in its home market, not the asking price of the one listing you found
• Card shops and marketplaces — pricing against what cards really sold for in Japan

How it works

Three sources, three separate infrastructures, read in parallel; paging inside each stays sequential and throttled to one request per 1.5 s. No browser, no login.

1. Yuyu-tei (遊々亭) — one request returns the whole catalog for the search word (no paging), giving setCode / cardNumber / rarity as typed fields
2. Yahoo! Auctions — closed auctions; an auction counts as sold only with bidCount >= 1, so ended-unsold buy-it-nows never reach a median
3. Mercari — sold-out listings only

Then the join runs locally: collector number first (114/098, 001/SV-P promo numbers canonicalized), then card name plus an explicit rarity. A title with neither is left unresolved rather than guessed. EX (2012) and ex (2023) are treated as different cards — case is the only thing separating ミュウツーEX from ミュウツーex, and conflating them put eleven ¥30,000 sales onto a ¥500 card.

Limits

• Japanese-language catalog. Card names must be in Japanese
• Coverage is 37–80 % per source and varies with how specific the listings are. It is always reported; it is never silently assumed to be 100 %
• Pokemon is the measured default. The collector-number match works for any game, but the Japanese set-name and series tables are Pokemon-tuned, so other games resolve fewer listings — visible in coverage.coverageRate
• Sampled statistics, not a full census: sources[].totalListingsFound is what the marketplace reports, sampledListings is what was actually read
• No image URLs and no description text are collected from the shop catalog (its terms of service forbid reuse of images and copy)

---

日本語: ポケモンカードが日本で実際にいくらで落札されたか(出品価格ではなく成約価格)を、ヤフオク・メルカリから型番(セット/カード番号/レアリティ)単位で集計し、PSA/BGS鑑定品と生カードを分けて返します。遊々亭の在庫あり提示価格との差も付きます。カード名1件 $0.02 の従量課金です。

中文: 输入日文卡名(如 リザードン,英文名在目录中查不到)即可获得宝可梦卡牌在日本的实际成交价(而非挂牌价),按卡编号/稀有度归集,并将 PSA/BGS 鉴定卡与裸卡分开统计。每个卡名 $0.02,按量计费。

If something goes wrong

• Wrong number or a failed run? Open a ticket on the Issues tab. I read every one and reply within 2 business days (Japan time).
• You never get a fake "empty" result. If the site can't be read, the run fails and says so.
• No results = no charge. You only pay for results you actually get.
• Checked every week. An automatic test runs this tool weekly; if the site changes, I fix it.
• Public pages only. No login, no personal data, and it goes easy on the site.

More tools by the same author

• Japan Used Camera & Lens — Dealer Price vs Sold Price
• Japan Figure & Gunpla — New Price vs Used Sold Price
• Used iPhone & Android Prices Japan — Dealer vs Sold
• Yahoo, Mercari & PayPay Japan — Same Item, 3 Prices
• Mercari & Yahoo Japan Demand — Sold vs Still Listed
• Mercari Japan Sold Prices — What Items Really Sell For
• Yahoo! Auctions Japan Sold Prices — Median, Range, Bids
• Mercari Japan Listings — One Row per Item, $0.0015 Each

All tools (Japan marketplaces, real estate, jobs, racing, prediction markets): https://apify.com/jpmarketdata

Disclaimer

Unofficial, independent tool — not affiliated with, endorsed by, or sponsored by Mercari, Yahoo! Auctions or Yuyu-tei. Product names and logos belong to their owners and only say where the data comes from. Data is read from public pages, for market research; check before you act on it.