Skip to content

How CardBrowse works

Where the prices, sales, and stock come from, how the indexes and Trends are built, and what the numbers don't tell you.

CardBrowse tracks Pokémon card prices, sales, shop stock, and what collectors are talking about, and updates once a day. This page explains every number on the site in one place.

Card prices and sales

  • Prices are TCGplayer Near Mint market prices, recorded daily (via PKMNPrices), through Nov 10.
  • Sales are real eBay sales matched to the exact card and printing, through Nov 7. Graded and raw sales are never mixed, each grade is kept separate, and outliers are excluded.
  • Listings are live eBay listings matched the same way, used for how many copies are for sale and what they’re asking.
  • Cards with too few price days, or fewer than 5 sales, are left out of a stat rather than guessed.

The Market page

  • The Market page leads with the CardBrowse Raw 100 (below).
  • Movers, sales, sold vs listed, and the segment scorecards come from 64 tracked cards chosen to cover vintage holos, modern chase cards, and the 30th Anniversary sets. The same basket every day, so changes reflect prices, not a changing list.
  • New releases (sets out for less than 90 days) have their own line and stay out of the headline numbers, because early prices mostly reflect release hype settling.

The CardBrowse indexes

CardBrowse publishes two indexes of the Pokémon card market. The CardBrowse Raw 100 follows valuable raw cards, priced daily. The CardBrowse Graded 100 follows the most traded high-grade slabs, priced weekly from real sales. Both started at 1,000. The Raw 100 launched Oct 24 with 100 members; The Graded 100 launched Oct 24 with 100 members.

The two indexes

  • Raw 100. 100 English cards with a Near Mint price of at least $100 and at least 6 raw sales on eBay in the last 90 days. Every card has the same weight. Prices are the TCGplayer Near Mint market price each day. Reviewed every month.
  • Graded 100. 100 English cards graded PSA 10, or PSA 9 for cards from the WotC era (1999 to mid-2003), where 10s rarely trade. A card needs a median sale of at least $750 and at least 4 sales at its grade in the last 90 days. Cards are weighted by their sales volume in dollars, capped at 5% each. Prices are the median eBay sale at the card’s grade over the last 4 weeks (at least 2 sales), and a card counts in a week only if it sold at least 6 times in the last 13 weeks. Reviewed every quarter.

Which cards qualify

  • English cards only, released at least 90 days before a review, so new sets settle first.
  • No jumbo, metal, error, or trainer-kit cards, and no deck exclusives.
  • At most 50 cards from one era, 8 from one set, and 2 printings of the same card, so no single set or era takes over.
  • Cards are ranked by sales volume in dollars (sales × median price). To limit churn, a new card enters only if it ranks in the top 90, and a member stays until it falls below 110.

Eras

Each index is also computed within three eras, from its members in that era:

  • WotC: 1999 to mid-2003.
  • Classic: Mid-2003 to 2019.
  • Modern: Sword & Shield onward.

The Graded 100’s WotC line is paused: too few WotC PSA 9 and 10 copies sell each week to track; back after the January review.

How the level is calculated

  • Each day (Raw 100) or week (Graded 100), every member’s price is compared with its price the period before. The index moves by the weighted average of those changes, and the moves are chained, so the line is continuous even when members change.
  • A card with no new price carries its last price for up to 7 days (Raw 100) or 1 week (Graded 100). After that it drops out of that period’s average until it trades again.
  • Outliers, Raw 100: a one-day move over 30% only counts when eBay sales around both dates agree with the old and the new price (within 35%), and no one-day move beyond 5× ever counts. This catches feed errors such as a card listed at $0.99 for a day.
  • Outliers, Graded 100: a card’s price can move at most a third up or a quarter down in a week, so a big repricing comes through over a few weeks instead of from one or two sales.
  • If any index moves more than 10% in a day (Raw 100) or 15% in a week (Graded 100), the update is held back and checked before it’s published.
  • History before launch is back-calculated using the launch members, and is shaded on the charts. It shows how today’s list would have moved, which flatters it slightly: cards were chosen partly because they trade well now.

Reviews

  • The Raw 100 is reviewed on the 1st of each month and the Graded 100 every quarter. Who entered and left, and why, is published on each index’s page.
  • Cards can be removed by hand between reviews for a clear reason (a confirmed data problem, a reprint that breaks the price series). Those removals are published too.

What the indexes don’t tell you

  • They track valuable, actively traded cards, not the whole hobby. Bulk, sealed product, Japanese cards, and other grades aren’t included.
  • Raw and graded prices are never mixed. A raw card’s price is TCGplayer’s Near Mint market price; a graded card’s is what copies at that grade actually sold for.
  • If a confirmed data error affects an index, it’s fixed, the affected history is recomputed, and the change is noted on the index page.

Trends shows what Pokémon card collectors are talking about, and checks it against what the market is actually doing. It updates once a day. The last update (Nov 10) covered 7,417 items from the last 28 days (350 news articles, 4,564 social media posts, 2,503 YouTube videos) across 246 communities, and found 101 public trends out of 598 tracked.

Where the data comes from

  • Social media. Public posts in the main Pokémon card communities, checked daily to see what collectors are discussing and how much attention each topic gets.
  • YouTube. New videos from about 150 Pokémon card channels covering openings, the market, grading, competitive play, news, and vintage. We compare each video’s views with that channel’s usual views at the same age, so a breakout stands out.
  • News. Pokémon card news from publisher feeds and news search. When several outlets carry the same story, it counts once.
  • Search. Monthly US Google search volumes for recent sets and the most-discussed cards (Google Keyword Planner, rounded ranges), the Google searches that bring people to CardBrowse (Search Console), and new Google and YouTube autocomplete suggestions.
  • Competitive play. Decklists from the top cut (top 8, or top 16 at 128+ players) of Standard tournaments with 32+ players on Limitless, compared fortnight to fortnight.
  • Release calendar. Upcoming English sets and products with their prerelease and release dates, kept by hand; dates not yet confirmed by The Pokémon Company are marked as reported.
  • The market. TCGplayer Near Mint prices, matched eBay sales, active eBay listing counts, and shop stock for 1,000 tracked cards, plus daily TCGplayer prices for booster boxes, Elite Trainer Boxes, and booster bundles.

How a trend is found

Each post, video, and article is read by AI to work out what it’s about (a restock, a price move, a grading result, a release, a scam) and which cards, sets, stores, or graders it names. Anything that names something specific is grouped with everything else about that same thing and event: every post about 30th Celebration restocks becomes one trend, whatever store or product it mentions. Posts that name nothing specific are grouped by subject.

Heat, confidence, and what gets shown

  • Heat combines how much a trend is mentioned this week, how many communities and sources it spans (each YouTube channel counts half, since many creators cover the same news), engagement, how unusual this week is compared with the trend’s usual share of the conversation, and market evidence. It’s on a fixed 0–100 scale, so it compares across days. A post that shares a video or article we already track counts once.
  • Rising and cooling compare this week’s mentions with what the trend’s usual share predicts for how much we collected this week, so a busier day of browsing or a newly added source isn’t a spike.
  • Confirmed means the conversation is backed by market evidence that fits it. The cards and sets that count are the ones the trend is about, named in at least a fifth of its posts. Price and hype trends: one of those cards selling at least 1.5× its usual pace on eBay, or moving 10% or more in a week compared with the typical card; or the set’s booster box, ETB, or bundle moving 8% or more. Hype and release trends only count a rise. Restock trends: a shop restock of the set’s sealed product, or eBay listings jumping 30%. Grading and scam trends are never “confirmed” by prices.
  • Selling faster than usual needs four full weeks of tracked sales for the card, so a card we only recently started tracking can’t look like a spike.
  • Early signal means conversation only, but from several communities or sources.
  • Stages turn this into a read: Building (conversation only so far), Confirmed (backed by market evidence), Priced in (confirmed, and the trend’s cards have already moved 10% or more over 30 days in the direction it points), and Cooling (mentions well below the trend’s usual share). Price trends point the way most of their posts say prices are going; hype and new-set trends point up.
  • A trend appears once it’s confirmed, an early signal, or reaches 5 mentions in a week, and it needs at least 2 mentions this week. Thinner trends are tracked quietly in case they grow. General genres like pack openings in general aren’t shown as trends.

What Trends doesn’t do

  • It doesn’t quote social media posts or show usernames. Posts appear as links with their titles.
  • Summaries describe what people are saying. They are claims, not verified facts, and never investment advice.
  • Off-topic stories that slip through are removed by hand.
  • Trends reflects the communities and channels we follow, which lean toward English-speaking, US collectors. A quiet week in our sources isn’t proof that nothing is happening, and the market evidence covers the cards we track, not every card.

Sealed product

  • Prices are TCGplayer market prices for English booster boxes, Elite Trainer Boxes (including Pokémon Center ETBs), and booster bundles (via PKMNPrices), recorded daily since Oct 19.
  • Per-pack prices use 36 packs a box, 9 an ETB (11 Pokémon Center), and 6 a bundle for Scarlet & Violet onward.
  • Shop prices for sealed product are covered under Shops.

New sets

New sets follows how each set’s most valuable cards move after release, so a new set can be compared with earlier sets at the same age.

  • Which sets: main English expansions and special sets since Scarlet & Violet (March 2023). Not Pokémon TCG Pocket (digital), promos, energy, McDonald’s, or reprint collections.
  • Which cards: each set’s 20 most valuable cards at TCGplayer Near Mint, one printing per card, $5 and up (fewer when fewer cards clear $5). New sets get their list at day 30; the list then stays fixed, so the line follows the same cards.
  • The line: equal weight, chained daily from the first day with prices. A card’s price carries over gaps of up to 5 days; single-day moves over 60% are left out as bad data.
  • Compared from day 30: the price source starts a new set’s history about three weeks after release, so sets are compared on their moves after day 30 (each line is 100 at day 30). That’s also when each set’s card list is fixed, and after the wildest stretch of opening hype.
  • Limits: the price source keeps 365 days, so sets released before October 2025 start partway through their life, and the comparison rests on a handful of sets for now. Sets tracked from a later list (rather than at day 30) lean towards cards that held up. How past sets moved doesn’t say what a new set will do.

Shops

The shop directory lists card shops across the US. For some of them, CardBrowse also reads what’s in stock, so their prices show on card and set pages.

  • How stock is read: once a day, from each shop’s public online store (the product feed or product pages its website already publishes), identifying as CardBrowseBot and following the shop’s robots.txt. Very large stores are read in full every couple of days and checked for changes in between. Shops can ask to be left out.
  • Matching: each listing is matched to the exact card, printing, and grade (PSA, BGS, CGC and others), or to a tracked booster box, Elite Trainer Box, or bundle. Listings that are unclear (an unnamed variant, a lot, a non-English card) aren’t shown. A sealed listing priced under half of market is held for review, because it’s more often another language’s box or a mislabel than a real deal.
  • Freshness: only stock seen in a read within the last 72 hours counts. Prices are the shop’s listed price, before shipping and tax.
  • Below market: a shop’s sealed price appears under “Below market at shops” when it’s in stock and at least 5% under the TCGplayer market price.
  • Some shop links can be affiliate links (see the privacy policy).
  • Run a shop and want it listed, corrected, or removed? Get in touch.

Monthly reports

Each month’s market report is built from CardBrowse’s stored data for that month: the indexes, TCGplayer Near Mint prices, eBay sales matched to the exact card and printing, sealed prices, shop stock, and Trends. The written summary is drafted from those numbers and edited by hand. Reports are published in the first days of the next month.

Searches and views

“Most searched” and similar attention numbers come from anonymous daily counts of searches, card views, and clicks to sellers on CardBrowse. Nothing that identifies a visitor is stored with them, and bots are filtered out. Details are in the privacy policy.

Limits and corrections

  • CardBrowse covers English Pokémon cards and sealed product; Japanese and other languages aren’t priced yet.
  • Prices and sales come from TCGplayer, eBay, and the shops we read. They describe what happened, not what a card will do next. Nothing here is investment advice.
  • If a price, match, trend, or index level looks wrong, tell us. Confirmed errors are fixed and the affected numbers recomputed.