TCGplayer market vs eBay sold: which price should you trust?

Trust TCGplayer market price for a raw, Near Mint, English card that sells at least a few times a week. Distrust any single number for a card that sells a few times a month. Use eBay sold listings, read by hand, for graded cards and for anything TCGplayer does not list. That is the whole answer; the rest of this guide explains why we built the site on the first number, where it breaks, and how to check a price yourself in a couple of minutes.

What TCGplayer market price is

TCGplayer's market price is a figure derived from recent completed sales of that card on TCGplayer's own marketplace, and it is published separately for each printing: Holofoil, Normal and Reverse Holofoil where they exist. TCGplayer does not say which condition the per-printing figure is for; in this site's check of five English Pokémon cards in October 2026 it matched TCGplayer's Near Mint price, so the site labels it Near Mint. It is not the lowest listing, it is not an ask, and it is not what a seller hopes to get. It is a summary of what buyers actually paid recently, which is the only thing a "price" should mean.

That gives it the property we care about most: it moves with real transactions rather than with listing behaviour. Where the site checked, it also matched TCGplayer's Near Mint price, which is why the site treats it as a Near Mint price.

Why it is our baseline

We pull TCGplayer market prices through TCGCSV, refreshed once a day around 20:00 UTC, for every card and sealed product in the catalog. Deal Score is a comparison against that number and nothing else; since October 8, 2026 a card listing is compared only when it is stated as Near Mint, the condition the number describes. The below-market price shown on card and product pages is computed from the same figure; a page shows none when the market is under $5, more than 10% above the highest TCGplayer listing, or more than 25% above the middle one. A seller's own "was $X" strike-through is ignored entirely. eBay listings, by contrast, are live asks, and an ask tells you what someone wants, not what the card is worth.

Using one baseline everywhere has a cost, which the rest of this guide is about, but it has a large benefit: the same card compared the same way in every listing, so a score of 80 on one card means what it means on another. The methodology page lists every rule.

The ten-sales-a-month rule

TCGplayer's own price-trend analysis only counts cards with at least ten sales a month. That is a useful confession from the source itself: below roughly ten sales a month, the market price is built on too few transactions to be treated as a fact. A card that sold three times last month, once to a buyer who overpaid, has a market price that is largely that buyer.

Most of a modern set is thin by this standard, and it does not matter much, because most of a modern set is worth cents. Here is how Pitch Black breaks down today.

As of September 27, 2026, the median card in the set is about 15 cents. An earlier version of this page gave the median as about a quarter, a figure that valued each card at its dearest printing rather than its standard one. The cards where a bad baseline could cost you money are the handful at $10 and above, and those are usually the liquid ones, since they are the ones people search for. The dangerous middle is the card at $8 to $30 that is not a chase card: rare enough to be thin, valuable enough to matter.

What the valuation apps get wrong

One comparison, published by the hobby site TCGinvest in July 2026, checked Collectr and PokeData against the median of seven cards' completed US sales over the previous 180 days: both ran a median of 6.5% above it, Collectr from 3% below to 25% above and PokeData from 33% below to 25% above. The same article cites TCGinvest's own wider study of 3,049 cards priced four ways, two of them on European prices: for cards with fewer than 31 completed sales, the highest and lowest of the four sat a median of 2.46 times apart, against 1.82 times for cards with more. The 6.5% is a small error. The gap between pricing methods is the larger one, and it is widest on cards without enough recent sales to pin a price. Any tool that prints a value for every card in a binder is printing guesses for most of them, and a "deal" that is 40% below a guess is not a deal. We do not use app valuations for anything, and we would suggest you do not compare eBay listings against them either.

What we do not have: eBay sold prices

eBay's completed and sold listings are the best public record of what graded cards sell for, and for many cards the deepest record of raw sales as well. We do not use them, because eBay does not expose sold-listing data through a public API. We could scrape it; we would rather not build the site on something that can be switched off, and we would rather not present numbers we cannot refresh reliably.

The main thing this costs us is graded cards. A PSA 10 does not have a TCGplayer market price, and comparing a slab to the raw Near Mint figure produces nonsense in both directions. So graded cards carry no Deal Score, and will not until there is a graded baseline we can maintain honestly. It also means that for a raw card TCGplayer does not list, or lists without a price, we have no baseline, no below-market price and no score. If a card page here shows no market price, that is usually why, not a bug.

Condition and the confidence score

Since the baseline is a Near Mint price, the Deal Score method scores an ungraded card listing only when its condition is stated as Near Mint. Deal Check reads eBay's structured condition descriptor first when the listing carries one; the search results the deal feed collects do not carry it, so since October 9, 2026 the feed fetches the full listing once for card listings whose condition reads as not stated or as Near Mint from the title and that would qualify if it said Near Mint, and otherwise reads words in the title, then eBay's generic condition label. Lightly Played, Moderately Played, Heavily Played and Damaged listings, and listings that state no condition at all, are shown without a Deal Score, because the site has no sourced price for a played copy. Until October 8, 2026 they were scored against fixed shares of the Near Mint market (Lightly Played 85%, Moderately Played 70%, Heavily Played 50%, Damaged 35%, and 90% when no condition was stated); those shares had no source.

Then the method gives every listing a confidence score from 0 to 1 that says how much to trust the comparison, and the Deal Score is scaled by it. Confidence starts from how the card was identified: a title with the set, the card number and the name matching a single card is the best case. It drops when the condition is read from the title rather than from eBay's descriptor, when shipping is calculated and had to be assumed at about $5, when TCGplayer's own price spread for the card is wide, when the card is worth under $10, when the seller is new or has feedback below 95%, and when the title suggests a quantity greater than one. Below 0.5 the listing is not called a deal at all. Confidence does not count sales, because the TCGplayer feed gives the site no sales counts; the nearest it comes is the wide-spread term above, and the less sure the comparison, the less the score is allowed to say.

How to sanity-check a price yourself

Take two minutes before any purchase over $20.

1. Open the card's page on this site and note the printing. A Reverse Holofoil and a Holofoil of the same card can have very different prices, and a listing title often does not say which it is. The photo does.
2. Look at the TCGplayer spread for that printing. When the lowest listing and the market price sit far apart, or when there are only a few listings, treat the market price as a range rather than a point.
3. Check eBay sold listings by hand for the same printing and condition, ignoring lots, graded copies and code cards. If the last several sales cluster around the TCGplayer figure, you have two independent sources agreeing. If they do not, the card is thin or the market has moved since the last daily refresh, and the sold prices are the fresher number.
4. Remember that an eBay seller pays roughly 13.25% plus $0.30 per order in fees, an effective 15 to 16%, so fixed-price asks cluster at or just under market. A listing 12% under market is our minimum to qualify as a deal; a liquid card at 40% under deserves a second look at its condition and its photos before you assume you have found a bargain.
5. Run the arithmetic yourself: landed price is item plus shipping, the baseline is the market price, and the discount is one minus landed over the market price; a card listing is scored only when it states Near Mint. The methodology page lists every rule and every reason confidence is reduced, so you can see which of the above is doing the work. Unless TCGplayer's market for the card is more than 10% above its highest listing or more than 25% above its middle listing, the card's page already shows the price that counts as 12% below market, and its eBay button opens search results capped under that price before shipping.

Bottom line

TCGplayer market price is the right baseline for raw Near Mint English cards because it is built from real recent sales, is published for each printing, matched TCGplayer's Near Mint price where the site checked, and can be refreshed daily for the whole catalog. It is a weak number for any card selling fewer than about ten times a month (on cards with fewer than 31 completed sales, one study found the highest and lowest of four pricing methods a median 2.46 times apart), and it is no number at all for graded cards, which is why slabs are not scored here. Deal Score scores a card listing against it only when the listing states Near Mint, the condition the site labels it, and compensates with a confidence score that shrinks any claim the data cannot support. For a card that matters to you, confirm the figure against eBay sold listings for the same printing and condition, and work through the landed-price arithmetic on the methodology page.