Quick answers about how Verdict works, where the numbers come from, and what to do when a search doesn’t return what you expect. If you don’t see your question here, send it over.
The Basics
What is Verdict?
Verdict is a trading-card decision tool. Type any card, TCG card, or non-sport card and get a read on what it’s worth right now — built from actual eBay sold listings, plus extra features for grading, flipping, lots, and collection tracking.
Where does the data come from?
Verdict draws from a stack of providers, each picked for what it does best:
- CardSight — primary catalog, pricing, and the vision provider for single-card AI Snap identification.
- Anthropic Claude — vision for the Grade Tool, Lot Calculator multi-card identification, and cert label scanning.
- GemRate — PSA / BGS / SGC / CGC / CSG grading population data and grade distributions.
- eBay sold listings — historical sales, cached in our database for long-term trend analysis.
- TCGplayer — ungraded (NM) market price for TCG cards in TCG mode.
- Apify — server-side eBay listing retrieval for Lot Calculator URL imports and as a fallback comp source.
- SerpAPI — fallback eBay sold-listing search when CardSight and the eBay API are both unavailable.
See How It Works for the full pipeline and how these sources combine into a single verdict. The full sub-processor disclosure lives on the Sub-Processors page.
Is it free?
Verdict has a Free tier with unlimited pricing search, Compare, and Flip Calc, plus a daily quota of AI Snap identifies, AI grades, and Lot Scout runs. Pro unlocks unlimited AI tools, cloud Collection and Watchlist sync, CSV export, and watchlist email alerts. Founders is a limited locked-in lifetime rate for the first 100 paid customers. See the Pricing page for current rates and the full feature list. You keep your existing data either way.
Do I need an account?
Anonymous use is fine for one-off pricing searches, Compare, and Flip Calc. You need an account to save a Collection, Watchlist, or Player Collection, to use the Grade Tool history, or to subscribe to Pro. Sign in with email + password or a one-tap magic link — no third-party (Google/Apple) login required.
How do I cancel?
Visit your account page and use Manage Subscription. That opens the Stripe Customer Portal, where cancellation is one click. You keep Pro features through the end of the current billing period, then your data stays on the Free tier. See the Refund Policy for details.
Pricing & Verdicts
How is Fair Market Value calculated?
It’s the median of recent eBay sold prices after a filter chain:
- Every word in your search must appear in the listing title (typo-tolerant fuzzy match).
- Sub-brand products that don’t match your query are removed (e.g., Chrome listings get stripped from a flagship Topps search).
- Non-card products (boxes, packs, lots, cases) are removed.
- Parallels and variants are separated when you’re searching a base card.
- Statistical outliers (wax boxes that slipped through, bizarre pricing) are removed using the IQR method.
The result is a median that represents a real, repeatable sale — not an inflated average pulled by one weird listing.
What’s “Weighted Recent”?
A version of the median that gives more weight to the most recent sales. If a card is trending up, Weighted Recent will show that before the straight median does.
What’s the Price Stability verdict (Stable / Volatile / Very Volatile)?
We look at how tightly the recent sale prices cluster. A card that sells every time between $45 and $55 is Stable. A card that sells at $40, $80, and $150 in the same week is Very Volatile — approach with caution.
Why are base cards sometimes grouped with a parallel?
When your query pins a specific insert (by card-number prefix like #SMLB-27or by the insert’s name like “Stars of the MLB”), we merge every matching listing into one bucket — because they’re all the same card. Splitting them into “Base vs Insert” would produce two prices for what is really one card.
Why do I sometimes see “No exact comps” with similar sales listed below?
For rare or obscure cards — a /5 auto, a one-of-one — eBay may have zero recent exact sales. Rather than leave you with a dead end, we walk a cascade of broader searches (drop the print run, drop the grade, drop the color modifier, etc.) and show actual eBay sales of similar variants. Each row is a real sold listing with a price, so you get a ballpark instead of nothing.
Pop Reports & Should I Grade
What’s the “Pop Report” block?
Whenever your search includes a grade (e.g., “PSA 10”), we pull a population report from GemRate that shows total graded population, gem rate, and 9+ rate across PSA, BGS, SGC, CGC, and CSG. Expand it for the full grader-by-grader breakdown. If population data isn’t loading, you may see a small “temporarily unavailable” notice — that’s a transient GemRate access issue and not a problem with your search.
I see a “Pop” button on some Collection / Watchlist rows. Why not all?
The Pop button only appears on rows whose card name includes a grade token (PSA, BGS, SGC, CGC, etc. + a number). Raw cards have no pop report to show — the button would be a dead end. Add the grade to your card’s name to enable it.
How does “Should I Grade” decide whether grading is worth it?
The Should I Grade panel on the Flip page combines CardSight’s pricing with GemRate’s grade-distribution probability data to compute a probability-weighted expected value:
EV = Σ (probability of grade i × net proceeds at grade i) − raw cost − grading fee
That math beats simply looking at the PSA 10 comp in isolation — a card with a small gem rate may still lose money on grading even when the gem-mint sale is huge.
The Grade Tool
How accurate is the AI grade estimate?
It’s an estimate, not a guarantee. The AI looks at centering, corners, edges, and surface in your photos and returns a projected grade band. It can’t feel the card, and it doesn’t know print defects only visible under specific lighting. Treat it as a data point, not a promise. See the disclaimer for details.
Why is adding the back of the card recommended?
For AI card identification (the “Snap a photo” button on the pricing page), the back is where the card number usually lives — dramatically improves accuracy on cards whose number isn’t clearly visible on the front. For the Grade Tool itself, we analyze both sides since centering and corner wear show up on both.
The auto-crop put the box in the wrong place — now what?
Tap where the correct corner should be and the nearest dot jumps there. Fine tune by dragging. If the auto-crop looks way off, tap Re-run with AI in the cropper header — that forces a full two-pass Claude Vision detection, which handles photos with messy backgrounds much better.
The Flip Calc
What’s the difference between the Flip page and the Grade & Flip panel on a search?
The Grade & Flip panel shows up automatically when you search a specific raw card that has a year and comps — it pulls PSA 9 and PSA 10 comps for that exact card and estimates grading profit. The Flip page lets you manually plug in a buy price and see how it compares to current comps at each grade tier.
Why doesn’t the Grade & Flip panel show on every search?
It hides when your query is too generic (no year specified) because the grade comps would be unreliable — you’d get PSA 10 “Chipper Jones 1/1” sales spanning 30 years of product. We only run the flip math when the card is pinned down enough to give meaningful comps.
Lot Calculator
How does the Lot Calculator price each card?
You paste a lot listing URL; we scrape the photos, AI-identify each card, and price each one through the same pricing stack as the main search — plus fallbacks: if a specific graded card has no comps, we try the raw equivalent, then a lower grade, then the base parallel, then similar cards. Each card is marked as an estimate when a fallback was used.
A lot card says “Not found” — why?
Every fallback in the chain came back with zero matches. Usually this means the AI misread the card (try cropping the photo tighter) or it’s an obscure parallel with no recent market. You can manually type the card to overwrite the AI’s guess.
Collection & Watchlist
What’s the arrow next to the price?
That’s the trend arrow — green up or red down — showing the delta versus the last time you checked this card’s price. Tap “Check” on a card to refresh it; the next check after that will show whether the value moved.
What’s the difference between Collection and Watchlist?
Collection is stuff you own — track purchase price, current value, and gain/loss. Watchlist is stuff you’re watching to buy — set a target price and we’ll tell you when the market drops into range.
Can I sort or filter my Collection / Watchlist?
Yes — the sort + filter controls appear above the card list once you have more than one card. Collection sorts by name, paid, current price, gain $, gain %, or last checked; filters narrow to priced, unpriced, winners, or losers. Watchlist sorts by target, current price, distance-to-target (deals first), or last checked; filters narrow to at-target, near (within 10%), above target, or unchecked. Sort and filter are session-only — they reset on refresh.
Why did “Check All” skip some of my cards?
Check All intentionally skips any card you’ve checked within the last 7 days. This protects our paid-API budget — a 100-card collection clicked twice a week would otherwise burn 200 calls on data that’s effectively unchanged. The skipped count is shown in the progress message. If you need fresher data for a specific card, the per-row Check button always refreshes immediately. If every card was checked recently, Check All will ask before re-burning the budget.
Accuracy & Edge Cases
The FMV looks too low / too high for a card I know well. Why?
Usually one of three things:
- Your search was too broad or too specific — try adding a year, removing a grade, or vice versa.
- A specific parallel is being filtered in or out. Check the Variants & Parallels panel and tap the correct one.
- There just aren’t many recent sales. Low-volume cards bounce around — look at the Sales Volume chart to see how much data we have.
Can I trust this for investment decisions?
No tool should be the only input to an investment decision. Verdict is a snapshot of recent market behavior. Use it as one data point alongside your own judgment. See the disclaimer.