# Methodology and editorial policy
Edited and reviewed by Mathis Grosjean, last reviewed on 2 November 2026. Who we are: [about Pepita](https://www.pepitacards.com/en/about). Contact: support@pepitacards.com.
Page: https://www.pepitacards.com/en/methodology
## Where do Pepita’s Pokémon card prices come from?
On 6 November 2026, Pepita prices 27,668 Pokémon cards and 1,750 sealed products from Cardmarket’s price guide, in euros. The price shown is the Cardmarket trend, only when sales back it (within ×3 of the lowest sold average). Otherwise a card may show its 30-day sold average, labelled as such, when its other averages agree (and TCGplayer, if priced). TCGplayer’s US market price comes via TCGdex, with TCGcsv as a fallback; PSA 10 prices come from PriceCharting. Prices are checked 4 times a day, and every priced answer is dated.
Sale prices, not buyback offers; condition, edition and language are not split. Source: Cardmarket, TCGplayer, PriceCharting · prices of 6 November 2026.
## Price data
Card prices come from Cardmarket’s public price guide, in euros, read through TCGdex, and straight from Cardmarket for cards TCGdex links to no product; sealed product prices come straight from Cardmarket’s price export. The price shown is the Cardmarket trend: for a card, the trend of its Cardmarket listing across all conditions and editions of that print, excluding reverse holo. Card identities (name, number, set, artwork) come from TCGdex. The English names of Japanese sets (VMAX Climax for VMAXクライマックス) come from Limitless TCG’s list of Japanese sets (limitlesstcg.com/cards/jp); a Japanese set it does not list keeps its Japanese name. PSA graded prices come from PriceCharting.
A trend is shown as a price only when sales back it: it must sit within ×3 of the lowest of Cardmarket’s sold averages, above or below. Further out, it is a broken reading, one old listing or a glitched trend: the card shows no single price and enters no ranking, and its page lists its Cardmarket readings instead (trend, average sale, sold averages), each labelled and dated with the day Cardmarket recorded them. One exception: when its 30-day sold average sits within ×3 of the average sale and of the 7-day average, and no more than ×5 above TCGplayer’s market price for the same print, the page answers with that 30-day sold average, shown as such, never as a trend and never in a ranking. For a sealed product, the average of the last week’s sales counts too.
When TCGplayer gives a market price for the same print of a card (normal or holo, never reverse holo), dated within 31 days of the Cardmarket trend and converted to euros at the ECB rate, the trend must not be more than ×5 above that price; beyond that, the card shows no price and its page lists its readings instead.
Every reading is added to our price history, which is never erased; before our own readings, it carries an archive of the Cardmarket trend. A past price is always quoted with the real date of its reading, and when there is no reading near a date, no price is quoted for it. Two dates serve as anchors: 30 December 2024, the archived reading that card and set pages compare today’s price with, and 17 June 2026, the oldest reading most sealed products share.
Rising and falling cards are ranked over 7 and 30 days, between two readings of the history. A card qualifies only at €5 or more today and €2.50 or more at the start of the window, with a trend backed by sales, and only when the last week’s sales average points the same way against the last month’s. Rises are capped at +100% and falls at −50%: beyond that, it is almost always a data error.
The week-by-week and month-by-month pages compare two exact readings (two Sundays, or the last full days of two months), with the same floors and the same caps (×2 and −50%). A fall is not named there when its start reading is a spike: with at least 5 readings in the 30 days before it, it must not sit more than ×3 above their median. Each table says how many of its rows sit within 1% of the cap, without changing the ranking. A day only counts when it is complete, judged against the days around it; a published week keeps its address, even if its list later goes away.
90-day rises start from the reading nearest the day 90 days earlier, at most 3 days away, and that reading must sit within 25% of the median of the readings in that window. The card must be worth €10 or more today, a price backed by sales, and €5 or more at the start, have at least 3 readings over the period and have moved in the last 30 days. Rises are capped at +200%, or ×3.
Sealed products are ranked from 17 June 2026: €20 or more today and €10 or more on that date, at least 30 days of history, a trend backed by sales, and rises capped at ×2.
A booster box’s price per pack divides its Cardmarket trend (or the median of several boxes) by 36, the packs in a standard English booster box; it is given only when every box in the figure is standard (no half, enhanced, 18-pack or Pokémon Center box) and its name states no other pack count. The booster box gaining the most value over 30 days is picked among the recent sets in the table (or, when none has a 30-day reading, among the 10 most expensive boxes) and measured from a reading dated exactly 30 days earlier, with the same floors and cap as the ranking above.
The PSA grading gain compares the PSA 10 price recorded on PriceCharting, converted to euros, with the Cardmarket trend of the same card, in the same finish on both sides (normal or holo; a holo card printed in a single finish is compared with its only print, even when PriceCharting does not call it holo). A card is compared only when PriceCharting counts at least 10 sales a year for that product, all grades together, and a raw-value band is shown only with at least 25 compared cards.
The site checks for new prices and is rebuilt 4 times a day, at 03:40, 09:40, 15:40 and 21:40 UTC. Every priced answer shows the date of its prices and links to this page.
## How shops are selected
Pepita tells apart a shop that is listed from a shop whose buyback is proven. A shop can help you prepare a sale without publishing its trade-in terms. In that case the page tells you to call, and never promises a buyback.
Our shop directory lists 898 physical shops in France. A shop is listed only when something concrete connects it to trading cards or Pokémon; a Google category is not enough. Unrelated businesses and online-only sellers are left out.
“Listed” means the shop may help you prepare a sale or ask about a trade-in. “Buys cards”, “takes bulk” or a percentage of the market price are shown only when one of the shop’s own pages says so clearly.
## Sources, dates and corrections
Addresses and opening hours come from the shops’ public business listings. Buy rates and trade-in terms are read from their websites. Every published figure keeps the link to its source and the date it was read. When something is missing, we do not fill it in.
Pages show the month of the last reading. Important data carries a “report an error” link. Before it is published, every correction is checked against a public source.
## Seller reviews
A shop’s Google rating mixes reviews from buyers and from sellers. Pepita keeps only the reviews whose author clearly says they sold, or tried to sell, cards to the shop. Buyer reviews, reviews about a shop employee who sells, and consignment do not count.
We never republish review text. Only overall figures appear, with the number of reviews and their dates. No average is shown with fewer than 3 seller reviews, and no recommendation with fewer than 5.
## Limits and responsibility
A shop can change its terms without notice. Call before you go, check the condition and the variant of your cards, and compare any offer with a recent market price. Pepita does not represent any shop and does not guarantee you will be offered a buyback.