Here is how to find card comps you can defend. Define the exact card first. Search by set and card number rather than player name. Pull sales from at least two sources. Discard hidden Best Offer prices, outlier auctions and unpaid orders. Separate qualified grades from clean ones. Then report a dated range, not a single number.
-
Write down the exact asset: year, set, card number, parallel, grader, grade, qualifier.
-
Search by set and card number, not by the player's name.
-
Pull from two or more sources and expect them to disagree.
-
Throw out the sales you cannot verify, and note why for each one.
-
Pull qualified, Authentic and Authentic Altered slabs out of the set.
-
Verify the slab in the sale, not the number in the title.
-
Check whether a deep comp set should even exist for this card.
-
Report a middle band and the date you pulled it.
The $900 card that does not exist
A collector types a rookie's name into eBay, filters to Sold, and sees results running from $180 to $2,400. The instinct is to treat that as volatility, split the difference, and land somewhere near $900. It is not volatility. It is four different cards and two bad sales stacked into one list.
For a modern rookie in particular, the comp set is wrecked less by fraud than by near misses. The base card, the Silver Prizm, Optic, Mosaic, Hoops and Fast Break versions, the numbered parallels, and outright customs and reprints listed with RC in the title all answer to the same player name. Filtering by card number and set rather than by player is the single highest-yield correction anyone can make to their process, and most collectors never make it.
Everything below is a funnel. Each step removes sales that were never comparable to begin with.
Step 1: Define the exact asset before you search
Write one line on paper or in a note: year, set, card number, parallel or base, grader, numeric grade, and whether the grade carries a qualifier. That line is the test every later sale has to pass. If you cannot state it, you are not pricing a card, you are pricing a name.
This matters most at the moment you are tempted to relax it. A comp set of two sales is uncomfortable, and the easiest way to get to eight is to quietly let a different parallel or a half grade in. That is the exact move that produces the $900 number.
Step 2: Search by card number and set, not by player
A name search returns every product that player appears on. A search built on the set and card number returns the asset you defined in step 1. The result count drops sharply and, more usefully, the price spread collapses, because most of the spread was never about this card.
Two mechanical notes on eBay itself. The public Sold filter covers roughly the last 90 days. Older sales still sit on eBay's servers but fall out of search, though sellers can request about two years of their own sold history through My eBay Data. And Sold is not the same as Completed: switching to Completed shows you what failed to sell, which is often the more honest signal about whether an asking price is real.
One caveat to check before you start. Multiple tooling vendors have reported that eBay moved signed-out sold and completed searches behind a sign-in wall in July 2026, with the signed-out route removed entirely the following month, leaving active listings unaffected. That claim comes from vendor blogs such as Scavio rather than from eBay, and we have not confirmed it against a primary source. If you find yourself redirected to a login page, that is the likely reason.
Step 3: Know what your source actually contains
Every comp source is a filtered view of a subset of the market. Knowing which subset is the difference between a comp and a rumor.
eBay's Trading Card Price Guide is built on eBay's own historical data from the past year, including transaction volume and the underlying transaction data, as eBay stated when it launched the trading cards experience in July 2021. That is a one year rolling window on one marketplace. Worth noting: on its current Seller Center page for the tool, eBay makes no accuracy claim at all, describing it only as a way to explore what cards may be worth. When the marketplace running the guide will not call it a valuation, neither should you.
PSA Auction Prices Realized is broader, aggregating over 5 million results from eBay, Goldin, Heritage and other houses, updated daily. It carries a caveat PSA prints itself: transcription and other errors may have occurred, and buyers should confirm a result against the actual sales catalog before making a decision. Treat a single startling APR line as a lead to verify, not a fact.
Card Ladder is vetted rather than raw, which is its whole value. More on its method in the next step.
130point queries eBay's own sold data by keyword and surfaces the accepted Best Offer prices eBay hides, and it also covers Goldin, PWCC, Heritage, MySlabs and Pristine over a window longer than eBay's 90 days. It is a search layer over real sales, not an independent price database, so it inherits eBay's problems while fixing one of them.
Use at least two. When they disagree, the disagreement is information about coverage, not a tie you need to break.
Step 4: The sales you throw out, and why
Best Offer sales where you only see the asking price. eBay's normal sold search hides the accepted amount on a Best Offer sale. A struck-through $90 may have changed hands at $60, and you have no way to know from that page. Worse, sellers on eBay's own community forums have reported the strikethrough and the Best Offer Accepted label disappearing entirely, which makes the asking price look like a completed sale price. That is why 130point exists. Either look the sale up there or drop it.
Auction results far above the cluster with a thin or odd bid history. You do not need to be paranoid about this, because both eBay and the data vendors treat it as a known problem. eBay's shill bidding policy prohibits bidding intended to artificially inflate a price, desirability or search standing, bars family, roommates, employees and online connections from bidding on a seller's items, notes the practice is illegal in many jurisdictions, and sets consequences ranging from listing removal to account suspension. Card Ladder, meanwhile, publishes its exclusion method: researchers study bidding patterns for obvious signs of shill bidding and route those auctions to a daily published purgatory list, apply the same treatment to extreme outlier sales inconsistent with reasonable market trends whether the sale was auction, Buy It Now or Best Offer, and exclude obvious scams from suspicious sellers outright. A professional data operation staffs a queue for this. An amateur is allowed to delete the row.
Cancelled and unpaid orders. These continue to appear under Sold. Nothing changed hands.
Sales where the shipping is the real price. eBay displays item price separately from shipping. A $0.99 sale with $20 shipping is a $21 sale, and in a raw list it reads as a dollar.
International listings. A different currency and a different shipping environment is a different market. Keep them out unless you are pricing for that market.
Step 5: The grade traps
PSA qualifiers mark a card that meets a grade in every respect but one: OC for off center, ST for staining, PD for print defect, MK for marks. PSA states there is no general pricing rule for qualified cards, because the discount depends on eye appeal and the difficulty of the issue. There is no percentage haircut you can apply. A PSA 8 (OC) is a different asset from a PSA 8, and it comes out of the comp set entirely.
Authentic and Authentic Altered slabs are not comps for anything numeric. PSA's grading standards explain why those holders exist: PSA will not grade cards showing evidence of trimming, re-coloring, restoration or other tampering. A card whose edge appears trimmed receives the N1 Evidence of Trimming designation, and at the grader's discretion it can be encapsulated as Authentic Altered when AA was listed as the desired minimum grade.
Half grades sit between numeric grades, not with them. And a card advertised as crossing over from another grader is a claim about a future outcome, not a grade. PSA's crossover service requires a stated minimum grade and returns the item in its original holder if it misses, with the full fee still charged. PSA has also suspended the service and removed the Minimum Grade option from the submission flow, so confirm the current status before you take a pending crossover seriously.
Step 6: Verify the slab, not the title
Run the certification number through PSA's cert lookup. Then read PSA's own warning in its buyer's security guide: criminals produce counterfeit labels using real certification numbers scraped from public sources. A cert number that returns a valid record proves that record exists, not that the slab in the photo is the slab PSA graded. The check that matters is comparing the card and label images PSA holds against that cert with the card actually in front of you, alongside the physical security features such as holograms, the tamper-evident case and the light-through-the-label check.
One more distinction that trips people up. eBay's Authenticity Guarantee for trading cards now triggers at $200 and above, reduced from $250, and covers raw cards as well as slabs from PSA, SGC, CGC Cards and Beckett. It verifies that the card or the sealed holder is authentic. It performs no grading or re-grading. An authenticated sale confirms the holder is genuine. It says nothing about whether the number on the label is right.
Step 7: How many comps do you need?
A working rule beats false precision. For a liquid modern rookie, aim for five or more clean sales inside the last 90 days. For a thin card, widen the time window before you widen the definition of the card. Relaxing the calendar costs you recency. Relaxing the definition costs you the whole analysis.
The PSA population report tells you which situation you are in, but only if you read it for what it is. PSA states plainly that pop figures represent cards PSA has graded, not all copies in existence, and that low grade cards are often never submitted, so the distribution skews high. Use pop to predict whether a deep comp set should exist at all. Do not use it to compute scarcity.
When several copies sell on the same day, average that day rather than quoting the most recent sale. That is not an improvisation: Card Ladder's last sold value is defined as the average value across the most recent day on which a card sold, recalculated nightly. The last sale is one draw. The day is a sample.
Step 8: Turn the survivors into a range
Discard the highest and lowest surviving sale. State the middle band. Note whether the band is drifting up or down across the window. That is the output. A band and a date, not a valuation.
If the band is wide, say it is wide. A $310 to $420 band on a card someone is asking $500 for is a complete answer, and a more useful one than a confident single number that quietly averaged a trimmed card and a shill auction.
Run the funnel yourself: a modern rookie worked example
The natural test case is a liquid modern rookie with a rich trap set, such as the 2018-19 Panini Prizm Luka Doncic base rookie in PSA 10. It has everything: a base card that shares a player name with Silver Prizm and every colored parallel, a heavy graded population, and enough volume that a real comp set should exist.
We are not printing figures for it here, and that is deliberate. No live comp data could be retrieved while this piece was written, because eBay's sold search and PSA Auction Prices Realized both refused automated access. Publishing a number pulled from memory or from an older article is exactly the failure this guide is about. A comp figure without a pull date is not a comp.
So run it yourself, and record what happens at each gate:
-
Raw count. Search the player name plus rookie and note how many sold results come back and what the full spread is.
-
After the card number filter. Re-search by set and card number for the base card only. Record the new count and spread. The gap between these two lines is the value of step 2.
-
After the Best Offer lookup. Run the same search on 130point and see how many struck-through prices resolve to a lower accepted amount. Note the largest single correction.
-
What you cut and why. Log every discarded sale with its reason: hidden accepted price, outlier auction, unpaid order, shipping-loaded price, international listing, qualified grade, Authentic or Authentic Altered holder.
-
The band. Drop the top and bottom of what survives, state the middle range and the direction of travel, and write the date of the pull next to it.
-
The test. Put the band beside the price the card is being asked at. That comparison, not the band alone, is the decision.
Keep that log. When someone challenges your number, the log is the argument.
The five-minute checklist
Write the exact card down first, including any qualifier. Search by set and card number, never by player name. Pull from two sources, one of which surfaces accepted Best Offer prices. Delete Best Offer sales you cannot resolve, auctions far above the cluster, cancelled and unpaid orders, shipping-loaded prices, and international listings. Pull qualified, Authentic and Authentic Altered slabs out entirely. Verify the cert images against the card, remembering that a valid cert number alone clears nothing. Check the pop report to see whether a deep comp set should exist. Average same-day sales. Drop the top and bottom, and state a band with the date attached.
The last thing worth being honest about: a comp range describes the last ninety days. It is a measurement, not a forecast. The point of doing it carefully is not to know what the card will be worth. It is to stop being wrong about what it is worth right now.
Related reading
Sources
-
eBay Community: Best Offer Accepted sold comps displaying weird
-
Card Ladder Help Center: Is Card Ladder's sales data reliable?
-
Card Ladder Help Center: Card Ladder Value, the Intersection of Player Indexes and Price Modeling
-
[PSA Grading Standards](https://www.psacard.com/gradingstandards)
-
Scavio: eBay Sold Listings Now Require a Login (secondary vendor report, unconfirmed against a primary source)
Note: 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.
