Grading arbitrage — buying raw cards, grading them, and selling the slabs for more than card-plus-fees — still works, but only as a screening business, not a volume business. The days of mailing PSA anything shiny and doubling your money ended with the 2021 backlog. What's left is a real but narrow edge for people who can grade condition themselves before paying someone $20 to confirm it.
The trade, defined by one equation
You're comparing the raw price against the probability-weighted slab value, minus all costs. Take the title example — a $30 raw card where the PSA 10 sells for $300 and the PSA 9 for $60. (Hypothetical but representative numbers; the structure is what matters.) Assume bulk-tier grading plus shipping runs about $25 per card all-in.
| Outcome | Probability | Slab value |
|---|---|---|
| PSA 10 | 40% | $300 |
| PSA 9 | 45% | $60 |
| PSA 8 or worse | 15% | $25 |
Expected slab value: 0.40 × 300 + 0.45 × 60 + 0.15 × 25 ≈ $151. After a ~13% selling fee: ~$131 net. Total cost: $55. Expected profit: ~$76 per card. Great trade — if those inputs are real.
Now watch it die from small input changes. Gem rate 25% instead of 40%: expected value drops to ~$106, net ~$92, profit ~$37. Gem rate 25% and a 10 that's only worth 5x raw ($150): expected value ~$68, net ~$59 — you're working for $4. The whole trade lives inside two numbers: your true gem rate and the 10-to-raw multiple. Everything else is logistics.
Rule one: the multiple picks the card
Never grade on absolute price; grade on the ratio. A $30 card with a 10x multiple beats a $454 card with a 1.8x multiple — and big modern chase cards often have compressed multiples precisely because their raw prices already assume gem potential. As of our July 2026 snapshot, cards like the 151 Charizard ex SIR (~$454 raw) or Prismatic's Umbreon ex (~$1,430 raw) trade rich partly because every buyer is pricing in the slab dream. The market pre-arbitrages the obvious candidates. The PSA 10 premium data shows where multiples are still wide — usually mid-tier cards from strict-condition sets, and vintage with honest surfaces.
Rule two: your gem rate is the business
Population-wide gem rates on modern Pokémon hover around a coin flip, but your gem rate is controllable through screening, and screening is the entire edge:
- Centering first. It's the most common gem-killer and the only defect fully visible in a listing photo. Measure it like a grader — the centering guide shows how — and reject anything near 60/40 before money moves.
- Surface under angled light. Print lines and holo scratches kill more 10s than whitening does on modern product.
- Edges and corners with a loupe. Especially dark-bordered cards, where factory silvering is rampant.
- Assume pack-fresh means nothing. A large share of cards are damaged before the wrapper opens — pack fresh isn't gem mint — which is also why "rip boxes, grade the hits" is a double-negative-EV plan. Most boxes lose money on raw EV alone (check any set on the Pack Value Calculator), and factory damage discounts the hits you do pull.
A disciplined screener can push gem rates well above the population average, because the population includes everyone who didn't look.
Where the edge died — and where it hides
The classic 2019–2021 trade — buy any raw chase at NM, grade at $10, triple your money — was killed by three things: grading fees that tripled, populations that exploded (every 10 you make competes with every 10 everyone else made), and a market that learned to price raws off slab values. Multiples compressed exactly where volume was easiest.
What's left:
- Mispriced raws in person — card shows and shops, where condition is inspectable and prices lag.
- Strict-set gems — sets with notorious centering or texture issues, where a true 10 is genuinely scarce and the multiple stays wide.
- Vintage condition reads — the deepest remaining edge, because judging 25-year-old surfaces is a skill few sellers have.
- Cross-grade and crack-out plays — narrower still, and fee-sensitive.
Also budget for the boring risks: 45–90 day turnarounds during which prices move, damage or grading disputes in transit, and the tax and fee stack on exit. Run the full worked math against your actual submission tier — the framework in should you grade modern cards applies card by card.
FAQ
What gem rate do I need to break even?
Solve the table above for your card: break-even gem rate ≈ (total cost − weighted non-gem value) ÷ (net 10 price − net 9 price). For typical modern cards with a 5x multiple and $25 all-in fees, you usually need 30–40% — before screening, many cards don't clear it.
Which grader should I use for arbitrage?
PSA, almost always — its 10s carry the deepest liquidity and the widest premiums. Other graders' slabs can be better value to buy, which is exactly why they're worse to sell.