How UltraBall prices Pokémon cards

Sales are public completed listings: eBay sold listings, Fanatics Collect, Goldin and Alt auction results, and a small number of other marketplaces. Each sale is recorded with its date, price, listing title, the grader and grade if it was a slab, and the listing’s own photo where the marketplace provides one. Asking prices are recorded separately and never treated as a sale.

Only English-language cards are covered. Japanese and other language cards are a different market and are left out of every count and price on the site.

A sale is matched to a card by its set, name and collector number. Before it counts, listings that don’t describe a single copy of that card are dropped: lots and bundles, several copies at once, proxies and custom cards, digital codes, non-English cards, and slabs listed under an ungraded price. A title that names a different card number, or a different set size, is treated as a different card and dropped too.

Some mismatches still get through. That is why every price rests on many sales rather than one, and why a card page shows the recent sales behind its number, so you can see the evidence rather than take our word for it.

1st Edition and Shadowless printings of the Wizards of the Coast era are separate entities with their own pages, their own sales and their own prices. The main card page is the Unlimited print. A sale is assigned to a printing from its listing title, and a printing is only priced when it has at least three of its own sales in the window.

Reverse holo sales are never mixed into a card’s price or sales list; they are a different market object and are left out until reverse holos have pages of their own. Where a modern card has several marketplace printings (holo, reverse holo, non-holo), the card page lists them separately.

An ungraded price is for a single English copy. Sales of played copies (listed as lightly, moderately or heavily played, or damaged) never set a near-mint figure, because they sell for less than a near-mint copy and would drag it below what a clean card fetches. A listing is what a seller hopes for; a sale is what a buyer paid. Alongside the price, the card page shows how many ungraded copies sold in the last 90 days in every condition, and the median of the near-mint ones, so the market price can be checked against like-for-like sales.

Condition matters: a played copy sells for less than the ungraded figure. The Rayquaza ☆ (Deoxys) page shows all of this in one place.

Where the headline ungraded price rests on sales stated as near mint, only those sales set it. A sale counts towards the condition its listing states and no other: a copy sold as damaged, heavily, moderately or lightly played goes to that condition’s own price, shown on the card page’s condition tabs, and never into a near-mint figure.

Many listings state no condition at all, and those are not treated as near mint. Measured on 1 October 2026 over about 190,000 ungraded sales of cards worth $10 or more in the previous 90 days: 28% of titles said near mint, mint or pack fresh and sold, in the middle, at our ungraded price; 52% stated no condition and sold 20% under it; titles stating light play sold 19% under, moderate play or just “played” 38% under, heavy play 53% under and damaged 62% under. A sale with no stated condition is counted in a card’s “all conditions” totals and nowhere else. In the lists of sold cards, an ungraded sale carries the condition its seller wrote (“sold as near mint”) or “condition not stated”: that is the seller’s word, not a condition anyone has checked.

Where a played condition has sold enough times, its price is the median of its own sales. Where it hasn’t, the price is an estimate placed between the conditions either side, and the card page says so. A worse condition is never priced above a better one. In a collection, an ungraded card is valued at the condition you record for it; one with no condition recorded is valued near mint, and the page says that too.

Older cards (sets up to 2016) can also carry a pack-fresh price: what the best ungraded copies sold for. It is our own measure, and it is built from sales alone: the top tenth of the card’s ungraded sales listed as near mint or better in the last 90 days, shown only when there are at least ten such sales, the figure is above near mint, and it sits under the card’s PSA 9 price. It is not taken from listings that say “pack fresh”: on older cards those sell, in the middle, at or under near mint, because the words are used loosely. Cards from 2017 on have no pack-fresh price, since nearly every copy is fresh from a pack and sells at near mint. In a collection, a card held as pack fresh is valued at that price where the card has one, and near mint where it doesn’t.

A PSA 10 and a PSA 9 are different markets and are priced separately, and PSA, CGC, BGS, SGC and TAG slabs are never mixed. Each graded price is for that exact grader and grade, and where we hold sales of it they are counted beside it. Some grades have no sale of their own and are filled in from a price guide; those carry limited confidence. A missing grade means too little to price it, not a worthless card. Sales counts on a grade row count only that grade.

A slab that sells often is priced at the middle of its recent sales, weighted towards the newest. A PSA 9 or 10 that sells only a few times a year is never priced below its last verified sale. That floor stops at slabs with thirty or more sales in 90 days: there one sale is one of many, and pricing a card at its single highest recent sale would overstate it.

Rankings such as the most valuable cards are ordered by ungraded value only. Graded values are shown alongside, but a single gem-mint slab never moves a card up the list.

A recorded sale that sits below a third of, or above 3 times, the market price for its grade is marked as an unusual price on the card page and is never used as a card’s last sale. These are usually a different card, a different printing, or a mislabelled listing. Record sales are only listed when they sit within half and twice the card’s price at that grade, so a listing ask recorded as a sale cannot become a record.

On price history, a single day more than 40% below or 67% above its neighbours is dropped, and a short run at 2.5× or more that snaps back is treated as the feed reporting another printing and removed.

The UltraBall Market Value is our estimate of what one copy in that condition sells for today. It is not the last sale, not the cheapest listing, and not an offer to buy or sell. Two cards with the same Market Value can be very different to sell, which is what the confidence and liquidity marks are for.

Why it is not a plain median of the sales

A card page shows one ungraded figure, the Market Value, and under it the sales it was built from. Those sales rarely agree with each other, and a plain median of them (every qualifying near-mint sale counted the same) would not be the same number. Where our own sales built the Market Value, it weighs them by how recent they are, what the listing said about condition, how the price was reached (an auction, a fixed price or an accepted offer) and how sure the match to this exact card and printing is, drops outliers, and follows the direction the market is moving. On a card with many recent sales that agree, the two would sit close. On a card whose sales are few, old, spread out or moving, the Market Value leans on the newest and most reliable sales. Any one sale on the page is one sale, not a price.

Confidence describes the evidence behind a price, not the card. It combines the source’s own confidence band with the number of sales in that exact condition over the last 90 days and how tightly they agree. Tap the mark on any price to see the facts behind it: qualifying sales, sales in the last 90 days, the typical range, and when the last sale was.

• High. Many recent comparable sales that agree closely.Many recent comparable sales that agree closely.
• Medium. Some recent sales, or sales that vary more than usual.Some recent sales, or sales that vary more than usual.
• Low. Few recent sales in this condition. Treat the figure as a rough guide.Few recent sales in this condition. Treat the figure as a rough guide.

Thirty or more sales in a condition in 90 days lifts a price to at least medium confidence; a hundred or more with a tight range lifts it to high. A price whose typical range is wider than the price itself is never high, and a graded price with fewer than three sales behind it is always low. When the last recorded sale is more than a year old the mark reads “stale evidence” instead.

A card can be worth $4,000 and still take months to sell. Liquidity says how easily a card sells in one condition. It is scored out of 100 from three inputs, all taken from that condition’s own sales: how many sold in the last 90 days (up to 60 points, on a curve where one a month scores about 12 and thirty a month scores 60), how recently the last one sold (up to 25 points), and how tightly sale prices agree (up to 15 points). Sixty-five and above is high, thirty-five to sixty-four medium, below that low.

• High liquidity. HighTrades often, at prices that agree. It counts sales, not how long a listing takes to sell.
• Medium liquidity. MediumTrades regularly, with some spread between sale prices.
• Low liquidity. LowFew recent sales in this condition. Selling may take time and the price is less certain.
• No recent sales. No salesNo recorded sales in this condition, so we cannot say how easily it sells.

We do not hold live listing counts or bid and ask prices, so those are not part of the score yet. Where a condition has no recorded sales, no score is shown rather than a guess.

Older history is recorded as monthly averages and recent history as daily prices. A four-week change compares today’s price with the same condition’s own line four weeks ago, and is only shown when both points sit in the daily record; a change never spans the join between the monthly and daily measures, and never crosses a fivefold jump caused by the feed reporting another printing. Every chart is smoothed the same way: daily prices are drawn as a 7-day average, run twice, so a single unusual day or a change of data source shows as a slope and not a cliff, and the last week of the line eases onto today’s price, so the figure under the chart is the price itself and not an average. A collection’s chart values each card at the condition or grade you hold it, and its last point is the total shown beside it.

The market page indexes are daily readings of a basket of the cards that change hands most, each counted at one grade and weighted by the square root of its dollar volume, capped at 5%. The cards are picked again on the first of each month, and each index page lists them and says how it is built. The long-run Pokémon card index is chain-linked monthly and rebased to 100.

Prices, sales and history are rebuilt once a day; the dates at the top of this page and in the footer say exactly when. Nothing on the site is real-time. A card that has not sold in over a year keeps its last price but its confidence mark reads “stale evidence”, and its last-sale date is shown so you can judge the age of the number yourself.

Prices are recorded in US dollars. Other currencies use the European Central Bank reference rate from Nov 11, 2026: a converted price is the US market value in your currency, not a local market price, which can differ because of shipping, fees and supply.

If a price looks wrong, first check the recent sales on the card page: they are the evidence behind the number, and a mismatched listing is the most common cause. To report a card whose sales are being matched wrongly, a missing printing, or a price that has gone stale, email hello@ultraball.com with the card page’s address. Reports are reviewed by hand and fixes ship with the next daily update. UltraBall does not buy or sell cards, and no price here is an offer; see about UltraBall for how the site is funded.