How far back does eBay's Sold filter go?

eBay's help page says completed listings results include only listings that sold in the previous 90 days. For older sales, eBay's Product research tool in Seller Hub covers the last 3 years.

How do I see what a Best Offer card actually sold for?

eBay says its Product research tool shows the actual price an item sold for, even if a Best Offer was accepted. Ball Card Genius also describes 130point as showing the actual sold price of Best Offer sales.

Is eBay Product research free?

According to eBay's news archive, Terapeak Product Research became free to all Seller Hub sellers from April 2021, with no standalone or Store subscription needed. Sourcing Insights still requires a qualifying Store subscription.

Should I use the average or the median of sold comps?

PureGrail's method uses the median of recent clean sales, because one outlier moves an average much more than a median. There is no single industry formula, and PriceCharting, for example, describes weighing recent sale, median, average and an age-weighted average.

Can I trust item specifics to filter parallels and grades?

Not on their own. Sellers fill in the Graded and Parallel/Variety fields, so they can be wrong or missing, and you should check photos and the card back against your exact card.

How to Find Sold Comps Cards Really Sold For

How to find sold comps cards really sold for: define the exact card (year, set, number, parallel, grader and grade), pull sales from eBay's Sold filter and its Product research tool, drop Best Offer sales with no visible accepted price, set aside lots and outliers, then take the median of recent clean matches.

Define the exact card before you search.
Pull sales from the right tools, and know how far back each one goes.
Remove Best Offer sales where you can't see the accepted price.
Set aside outliers that look like shill bids.
Remove wrong parallels and mismatched grades.
Put every sale on the same basis: shipping, premiums, no lots.
Take the median of the most recent clean sales and check it against active listings.

The goal is not to collect as many comps as you can. It is to end up with a small set of sales that match your card exactly, show the price actually paid, and are recent enough to matter. Five clean comps tell you more than thirty dirty ones. This is a refresh of our sold-comps guide, rebuilt around what eBay's own documentation says its tools do and don't show.

Step 1: Define the exact card before you search

Write down the card before you type anything into a search bar:

Year and set (including the specific product line, not just the brand)
Card number
Parallel or color, plus serial numbering if it has any
Autograph, relic or patch, if any
Grader and grade, or an honest raw condition if it is ungraded

Any sale that doesn't match on every line comes out. This sounds obvious, but most bad comps come from skipping this step. Someone searches a player name and a year, sees a cluster of prices, and anchors on a number that mixes base cards, parallels and three different graders.

Step 2: Know where the sales data lives

eBay's Sold filter: 90 days

The public Sold or Completed filter on eBay search is where most collectors start. Its window is short. According to eBay's Product research help page, "eBay completed listings results include only listings that have recently ended sold in the previous 90 days." For a card that sells weekly, that's plenty. For a scarce card, 90 days may show one sale or none.

eBay Product research: 3 years

eBay's Product research tool (formerly Terapeak) sits under the Research tab in Seller Hub. eBay's help content says it "gives you access to the last 3 years of eBay sales data for millions of items." According to eBay's news archive, "From April 2021, eBay makes Terapeak Product Research available for free to all Seller Hub sellers," and sellers "will no longer need a standalone or Store subscription to access it." (Sourcing Insights, a separate feature, still needs a qualifying Store subscription.)

The same help content lists the metrics Product research shows: average sold prices, average shipping costs and terms, number of sellers who sold the item, sell-through rates, upward or downward trends, sales trends and sold price range. You can filter by category, item condition, brand, buyer and seller locations, product identifiers, sale price ranges, listing formats and eBay marketplaces. The listing-format and price-range filters are the useful ones for comps.

130point

130point is the best-known free third-party option. The hobby site Ball Card Genius describes it as "a website separate from eBay that will tell you the actual sold price for items that were sold at best offer." We weren't able to review 130point's own documentation, so we describe it only through that attribution.

Step 3: Deal with Best Offer distortion

This is the most common way a sold price on screen misleads you. When a seller accepts a Best Offer, the public sold results may show the listing's asking price, not the price paid. A card listed at $150 that sold for $110 can look like a $150 comp.

Overhead view of sleeved trading cards grouped under a lamp's light on dark felt, with one card pushed alone into the shadowed edge of the mat beside a loupe.

eBay's help page says Product research results "include item sales with the actual price that the item sold for, even if a Best Offer was accepted." That makes Product research the main fix. Members of an eBay Community thread point to the same tool for this.

Don't rely on visual cues in the public view to flag these sales. In another eBay Community thread, users reported that the sold section sometimes no longer showed a strikethrough price when a Best Offer was accepted. No eBay staff replied, so treat that as user reports. The practical point holds anyway: you can't count on the public sold view to tell you which sales were offers.

Rule: if you can't see the accepted price, drop the comp. At most, treat it as a ceiling (the card sold for this much or less), never as a data point in your average.

Step 4: Set aside outliers that look like shill bids

eBay's shill bidding policy defines it plainly: "Shill bidding is when someone bids on an item to artificially increase its price, desirability, or search standing." The policy lists possible actions including "Removing the listing or other content, issuing a warning, restricting activity or account suspension," and notes that shill bidding is illegal in many places.

You usually can't prove a sale was shilled, and you shouldn't accuse a specific seller from a price chart. You don't need proof to keep a suspicious sale out of your average. Set a sale aside when you see:

One sale far above a tight cluster of otherwise similar sales, with nothing near it before or after
An auction result above every Best Offer and Buy It Now sale around it
A big sale that is never followed by similar sales, or where the same card seems to reappear for sale soon after

"Set aside" means you keep it in view but don't average it in. If later sales confirm the higher level, it rejoins the data.

Big headline sales can be disputed too. In August 2021, ICv2 reported that eBay said "individuals associated with a trading card seller, PWCC, have engaged in 'shill bidding'" and removed PWCC's auctions. ICv2 reported PWCC had over 17,000 active auctions, all gone by the end of the day. PWCC responded: "PWCC has never engaged in nor condoned the practice of shill bidding or manipulating the card market."

In 2025, Sports Collectors Digest reported that Alt sued PWCC in New York State Supreme Court, alleging a years-long shill bidding scheme. SCD quotes the complaint: "Alt claims that its own 'independent analysis' showed that PWCC had access to customers' private maximum bids, allowing it to drive up the price." SCD reports Alt said it spent "more than $10.7 million on trading cards between 2021-23," and the sales cited include a $528,000 2017 Panini Prizm Patrick Mahomes rookie (up from $16,500 in 2019) and a $3.12 million Luka Doncic card. A PWCC spokesperson said: "Alt has filed a completely baseless lawsuit against PWCC that lacks any substantive evidence."

These are allegations, not findings, and we take no position on them. The lesson for comps is narrower: a widely reported sale is not automatically a clean data point.

Step 5: Remove wrong parallels and mismatched grades

This is the most common error after Best Offers. Watch for:

Macro close-up of two trading card corners held together under a loupe, their foil textures catching raking light at different angles to show a visible difference in pattern.

Base cards mixed with refractors, Prizm colors or other parallels
Numbered parallels mixed with unnumbered ones, or different print runs of the same color family
PSA 10 mixed with PSA 9, or PSA mixed with other graders at the "same" number
Raw cards titled "gem mint" or "PSA 10?" that are ungraded

eBay's item specifics can help you filter, but they aren't reliable on their own. According to CardLines' eBay listing guide, the Graded item specific is required. For graded cards, sellers enter the grader and certification number, and refractor or Prizm-type details go in the Parallel/Variety field. Sellers fill these in, so they can be wrong or blank. Check the photos and the card back against the title and against your card.

Never mix graders or grades in one average. A PSA 9 and a different grader's 9 are separate markets. A 9 and a 10 can be very different markets.

Step 6: Put every sale on the same basis

Shipping: decide whether you are comparing prices with or without shipping, and apply it to every sale. Price guides differ here. PriceCharting, for example, says its prices exclude shipping and fees.
Auction-house results: a hammer price may or may not include a buyer's premium. Check the house's current terms before putting an auction-house result next to an eBay sale.
Lots and bundles: exclude them. You can't separate the price of one card from a three-card lot.

Step 7: Weight recent sales (the PureGrail method)

There's no single industry formula for this. What follows is PureGrail's editorial method, not a standard:

Build from the last few clean sales, roughly five to eight if the card trades often enough.
Use the median, not the mean. One bad sale moves an average a lot and a median very little.
Give the newest sales more weight. If the last three clean sales all sit below the older ones, the market has probably moved, and the median of the full set will lag.
Check against current asking prices. Ball Card Genius makes this point: if a card has sat listed at $10 for weeks, recent $20 comps may be outdated.

For comparison, PriceCharting describes its own approach this way: "Our algorithm will consider things like most recent sale price, median price, average price, age weighted average price, and much more. Our algorithm takes outliers into account and the date when sales occurred." Its FAQ adds that "Listings that sold for an abnormally high or low price are automatically flagged for review by admin to determine if they are legitimate sales," and users can report historic sales for review. That FAQ wording is centered on video games, so read it as one price guide's method, not proof of how any card price you see was built.

Worked example (hypothetical numbers)

This is an illustration. The card and every price below are made up to show the method, not to describe any real market.

You're pricing a PSA 9 base rookie of a hypothetical player. A raw search returns ten sales, all shown before shipping:

| # | Sale | Shown price | Decision | | --- | --- | --- | --- | | 1 | PSA 9 base, auction | $120 | Keep | | 2 | PSA 9 base, Best Offer accepted, accepted price not visible | $135 | Drop (ceiling only) | | 3 | PSA 9 base, Buy It Now | $118 | Keep | | 4 | PSA 9 refractor | $240 | Drop (wrong parallel) | | 5 | PSA 10 base | $310 | Drop (wrong grade) | | 6 | PSA 9 base, auction | $125 | Keep | | 7 | Lot of three slabs including this card | $95 | Drop (lot) | | 8 | PSA 9 base, auction, no similar sale before or after | $410 | Set aside (outlier) | | 9 | Different grader, 9, base | $112 | Drop (wrong grader) | | 10 | PSA 9 base, Buy It Now | $122 | Keep |

The average of all ten shown prices is $178.70. That number describes no real card.

Four clean comps remain: $118, $120, $122 and $125. The median is $121. If you had left the $410 outlier in, the mean of those five would jump to $179, while the median would only move to $122. That gap is why we use the median.

Last, check active listings. If several PSA 9 copies are sitting unsold at $105 to $110, the market for your card is probably below $121, not at it.

When comps are thin

Low-population and high-end cards may sell once or twice a year. The 90-day Sold filter won't help much. What to do:

Widen the window with Product research's 3 years of data, and be clear that older sales describe an older market.
Look at neighboring grades and close parallels for direction only. A PSA 8 sale can tell you whether a 9 is likely to be above or below it, not what the 9 is worth.
State the uncertainty. "Last clean sale was 14 months ago at X, and nothing comparable has traded since" is more honest and more useful than a forced number.

Using the result: grade, sell or hold

A clean comp is an input, not a decision. If you're deciding whether to grade a raw card or sell it as is, plug your raw and graded comps into PureGrail's card grading break-even calculator. It compares the incremental net proceeds of grading against selling raw, using the costs you enter: grading fees, shipping, marketplace fees and your comp prices. It does not predict what grade your card will get, and it does not supply current prices or forecast returns. The output is only as good as the comps you put in, which is why the steps above matter.

Quick checklist

Exact match: year, set, number, parallel, serial numbering, auto or relic
Accepted price visible (Product research, or a third-party tool for Best Offers)
No lots or bundles
Outliers set aside, not averaged in
Same grader and same grade only
Same shipping basis; auction premiums checked
Median of recent clean sales, with the newest weighted more
Sanity check against current active listings

Sources

This article contains AI-assisted content and has been reviewed in our editorial workflow.

DISCLAIMER: PureGrail articles are for informational and entertainment purposes only. Nothing on this site constitutes financial, investment, or legal advice. Collectibles are speculative assets and values can decrease significantly. Always conduct your own research before buying or selling. Past price performance does not indicate future results.

How to Find Sold Comps That Reflect a Card's Real Market

Step 1: Define the exact card before you search