Grading Opportunities

A live feed for signed-in Trader-plan members: raw cards CardTrack has seen listed on eBay where grading looks worth it — weighted across every grade outcome, not just the PSA 10 dream, and net of the card, the grading fee and the sale.

Gem rates come from PSA's published population (CGC where PSA is thin). Population is cumulative and submission-biased — nicer copies get submitted and resubmissions inflate high grades — so these are approximations of a historical rate, not probabilities for the card in the photograph. Cards the seller hasn't described as Near Mint or Lightly Played are excluded entirely. Market analysis, not investment advice.

How the feed decides

Every row starts as a raw eBay listing that a Shop Scanner run has already matched to a specific card and printing. The engine then asks one question: if this copy were bought at the listed price, graded, and the slab sold, what would be left? The answer is an expectation, not a best case — the PSA 10 price weighted by how often the card gems, the PSA 9 price weighted by how often it lands a 9, and the sub-9 prices from the card's own graded ladder weighted by the rest — minus the purchase price, the grading fee for the grade actually received, and the fees and postage on the eventual sale.

Only listings whose expected net clears $25 appear here, ranked by that figure. A row shows the buy price, the grading cost, the gem rate with the sample it came from in the PSA population report, and the seller's stated condition, so a positive number can be read back to its inputs rather than taken on trust. The ROI figure is the expected net divided by everything paid before the slab is sold.

Two rules are stricter than the arithmetic. The gem rate is only allowed to speak for a copy the seller describes as Near Mint (the condition guide sets out what that means); a Lightly Played copy is kept but flagged, because the population's gem rate assumes a typical submitted copy and a lightly played one gems less often than that. Anything described as Moderately Played or worse, or not described at all, is left out entirely — a heavily played card grades roughly nothing at PSA 10 whatever its population says, and a listing with no condition gives the engine nothing to reason from.

The rule of thumb

Strip the machinery away and the test is the one every grading decision comes down to: grading pays when the odds-weighted graded value beats the price of the card plus the fee, with room to spare. The multiple between raw and PSA 10 — the PSA 10 premium — is the number people quote, and it is the least useful one, because it is only collected on the submissions that gem. A card that gems one time in twenty needs its PSA 10 to sell for at least twenty times the grading cost above the raw price just to break even, if the other nineteen submissions come back worth only raw.

The safer version of the rule is to ask which grade first pays for itself. When the PSA 9 price already covers the buy price and the fee, the play survives the most common near miss and the 10 is upside; when only a 10 covers it, the whole return rests on the hardest grade to get. The feed's expected-net figure blends both cases, but the break-even grade is the quicker sanity check when a row looks too good — the grading ROI calculator reports it for any card.

What the catalogue says

The feed only lists cards someone is currently selling, so its size moves with the market. The catalogue behind it is steadier: of the 2,771 Pokémon cards CardTrack prices at $20 or more raw with a coherent PSA 10 price above that, 2,591 (94%) would clear PSA's current fee and shipping if they graded a 10.

Odds narrow that a lot. Among the 2,448 of those with at least 100 PSA-graded copies behind the gem rate, the median gem rate is 11.5%, and under the conservative assumption that anything short of a 10 is worth only the raw card, grading pays on 978 — 40%. The engine behind the feed prices the 9 and the sub-9 grades from the card's ladder rather than at raw, so its own count is higher, but the shape is the same: most cards clear at a 10 and a minority clear on the odds.

By era, the cards that pay on the odds are led by Modern (2011–2019) with 391, Vintage (1999–2003) with 241, Classic (2004–2010) with 177 and Current (2020–) with 169. Older sets earn their place through scarcity of clean copies; recent sets through chase cards whose PSA 10 premium is wide enough to absorb a modest gem rate. The best cards to grade list can be filtered by era to see them.

What grading costs, and how long it takes

The grading cost on each row is PSA's fee for the tier that covers the graded value plus round-trip insured shipping — $20 flat plus 1% of declared value — at fees as of 2026-09-26. PSA's cheapest bookable tier, Standard, is $59.99 per card for slabs up to $1,000 and quotes 90–100 business days; the fee steps up as the declared value crosses each tier's cap, so a card whose PSA 10 sits just above a cap costs noticeably more to grade than one just below it. The grading cost guide has the full PSA, CGC and BGS tables.

Turnaround is part of the cost. At 90–100 business days plus postage, the slab comes back into a market that has had roughly 20 weeks to move — the biggest movers list shows how far that can be — and a listing that looked cheap because a card was spiking may look ordinary by then. The feed cannot see the future price, so a row is only as good as the assumption that the graded price holds.

Why some listings are refused

A missing row is often a refusal rather than a miss. The engine returns nothing, not a guess, when the card has fewer than 10 graded copies in the PSA and CGC population reports, because a gem rate built on a handful of submissions is arithmetic rather than evidence; when the card's PSA 10 price is incoherent against its raw price, which is a sign the graded figure is stale or polluted; and when the condition gate above says no. Rows built on fewer than 100 graded copies carry a caveat, since the rate is real but thin.

Common mistakes

Buying the gem rate rather than the card. The population rate describes copies other people chose to submit, most of them pre-screened. A listing photo showing edge whitening should override a 30% gem rate, not be averaged with it — which is exactly why the feed keeps the seller's condition on every row, and why the AI card grader is worth running on the photo before buying.

Forgetting the second sale. Expected net here already deducts marketplace fees and postage on selling the slab, which is what makes it comparable to the buy price; a calculation that stops at graded value minus fee overstates the return, and on a $200 slab the eBay fee calculator shows the difference is real money.

Grading everything that clears. A positive expected net across a hundred submissions is a portfolio, and a single submission is one draw from it — the typical outcome on many rows is a 9 or an 8 that returns less than the buy price. Size the batch so that the losing draws are affordable; the submission planner totals a whole batch before it is sent.

FAQ