Is grading a Pokemon card worth it? The break-even math

I get this question every week. Sometimes from a friend with one Charizard. Sometimes from a dealer with sixteen slabs they're considering crossing over to a different grader. The framework is the same.

The full answer is one equation, three input variables, and four worked examples. Here it is.

The equation

Expected_Value_Graded
  = (P_top_grade × Sale_top_grade)
  + (P_one_below × Sale_one_below)
  + (P_two_below × Sale_two_below)
  - Submission_Fee
  - Shipping_Round_Trip
  - Time_Cost_of_Capital

Compare that to the raw sale price. If Expected_Value_Graded > Raw_Sale_Price + Buffer, submit. Otherwise sell raw.

The buffer is the dealer judgment piece. Mine is 15% — I want a 15% margin on top of the raw price to justify the wait, the risk of damage in transit, and the opportunity cost of capital tied up.

The three input variables (in order of importance)

Variable 1 — Probability of grading top tier (and one below)

This is the hardest call. Grading services are not consistent across submissions, but a single dealer's hit rate on similar cards is reasonably predictable.

My rough hit rates on Pokemon cards in NM/M condition (eyeballed, by me, on cards I've personally selected):

These are my rates with my sourcing and my condition eye. Yours will differ.

The dealer-error trap: assuming you'll grade higher than the data says. If your last 30 submissions show a real P(PSA 10) of 35%, planning around 50% is going to disappoint you.

Track your hit rates. Adjust your math against your actual rates, not your hoped-for ones.

Variable 2 — Sale prices at each grade

This is the easier variable. Pull comp data:

• PSA 10 sale price: median of last 10 PSA 10 sold on eBay
• PSA 9 sale price: median of last 10 PSA 9 sold on eBay
• PSA 8 sale price: median of last 10 PSA 8 sold on eBay
• Raw NM sale price: median of last 10 raw NM sold on eBay

The vault's auto-pricing engine surfaces all four of these on every gradable card. If you're hand-pricing, the eBay sold filter is your friend.

Variable 3 — Fees + shipping + time cost

Easy:

• PSA Value tier: $25/card
• PSA Express: ~$300/card
• CGC Bulk: $14/card (10-card minimum)
• Shipping round trip: ~$15-20 per submission (split across all cards in the submission)
• Time cost of capital: I use $5 per $100 of expected value, per month tied up. For a 60-day Value-tier turnaround that's roughly $10 per $100. Adjust to your own opportunity cost.

Four worked examples from a typical submission batch

The four examples below are representative of the decisions a dealer makes on every submission — drawn from the actual math I run, with card identities and probability assumptions tuned to make the framework legible. Use the structure on your own cards; your hit rates and comps will differ.

Example 1 — Crown Zenith GG39 Rayquaza VMAX Alt Art, raw NM

Raw NM sale price: $160 (eBay 30-day median). PSA 10 sale price: $1,400 (eBay 30-day median). PSA 9 sale price: $440. PSA 8 sale price: $200.

My hit rate on Crown Zenith fresh-pulled, careful-handling cards: P(10) = 55%, P(9) = 35%, P(8) = 10%.

EV_graded = (0.55 × 1400) + (0.35 × 440) + (0.10 × 200)
          = 770 + 154 + 20
          = 944
EV_net    = 944 − 25 (PSA Value) − 17 (shipping share) − 56 (time cost)
          = 846

Compare: raw sale = $160. Net graded EV = $846. Margin = 5.3x.

Easy submit. I sent it.

Example 2 — 2002 Japanese Neo Destiny Shining Tyranitar #110, raw LP

Raw LP sale price: $80. PSA 10 sale price: $1,600 (sample size: 3 in the last 90 days, high variance). PSA 9 sale price: $400. PSA 8 sale price: $180. PSA 7 sale price: $90.

My hit rate on WOTC-era Japanese, used-binder source, LP condition: P(10) = 3%, P(9) = 20%, P(8) = 35%, P(7) = 30%, P(6 or lower) = 12%.

(Note: LP cards rarely grade 10. The "S" on the gradable scale starts at 9 for most LP cards.)

EV_graded = (0.03 × 1600) + (0.20 × 400) + (0.35 × 180) + (0.30 × 90) + (0.12 × 40)
          = 48 + 80 + 63 + 27 + 4.80
          = 222.80
EV_net    = 222.80 − 25 − 17 − 13
          = 167.80

Compare: raw sale = $80. Net graded EV = $168. Margin = 2.1x.

Submit, but it's a closer call. The variance on PSA 10 sale price is high; if the next PSA 10 sells at $1,000 instead of $1,600 the math gets thinner. I sent it but I marked it "if it comes back PSA 8 or below I sell, don't list at the long tail."

Example 3 — 2024 Stellar Crown Latias ex 217/142 (Alt Art), raw NM

Raw NM sale price: $52. PSA 10 sale price: $115. PSA 9 sale price: $58. PSA 8 sale price: $40.

My hit rate on Stellar Crown careful-handling: P(10) = 60%, P(9) = 32%, P(8) = 8%.

EV_graded = (0.60 × 115) + (0.32 × 58) + (0.08 × 40)
          = 69 + 18.56 + 3.20
          = 90.76
EV_net    = 90.76 − 25 − 17 − 5
          = 43.76

Compare: raw sale = $52. Net graded EV = $44.

Do not submit. The card is not worth grading at this comp tier. Sell raw at $52. I held it back and listed raw the same day.

This is the result most dealers get wrong. They see "this card is worth $115 in PSA 10" and they assume submission is automatic. The math says otherwise — at this comp gap, the grading service is taking your margin.

Example 4 — 1999 Base Set Charizard #4 1st Edition, raw NM

Raw NM sale price: $2,800 (eBay 30-day median). PSA 10 sale price: $34,000. PSA 9 sale price: $8,500. PSA 8 sale price: $4,200. PSA 7 sale price: $2,800.

My hit rate on WOTC-era English from a trusted condition source (this card came from a careful long-term collector): P(10) = 5%, P(9) = 25%, P(8) = 40%, P(7) = 25%, P(6 or lower) = 5%.

EV_graded = (0.05 × 34000) + (0.25 × 8500) + (0.40 × 4200) + (0.25 × 2800) + (0.05 × 1500)
          = 1700 + 2125 + 1680 + 700 + 75
          = 6280
EV_net    = 6280 − 200 (PSA Regular tier; >$1k declared value) − 30 − 280 (time)
          = 5770

Compare: raw sale = $2,800. Net graded EV = $5,770. Margin = 2.1x.

Submit, but go up a tier. I used PSA Regular tier ($200) instead of Value because the declared value above $1k requires it. The 2x return justifies the fee bump.

The pattern

Looking at all four examples:

• Big-gap cards (Crown Zenith GG39, vintage Charizard): always submit. The grading service captures a fraction of the gap; you keep most of it.
• Marginal-gap cards (Japanese vintage LP): submit if the variance is acceptable, hold back if it isn't.
• Small-gap cards (Stellar Crown Latias ex): do not submit. The grading fee eats the margin. Sell raw.

The rule of thumb that emerges:

If PSA_10_sale_price < 3 × raw_sale_price + fees, do not submit.

That's the "is this card worth grading at all" gate before any of the math.

What this means for your inventory

Run this math on every card above $30 that's gradeable. The vault does it automatically — pulls comp data for raw / PSA 8 / PSA 9 / PSA 10, applies your hit rate (you can override the default), and surfaces a "submit / hold / sell raw" recommendation per card.

If you're hand-doing this, expect 90 seconds per card. Worth the time on cards above $50; not worth it below.

One more thing

The math above is for PSA. Sub-in CGC fees ($14 instead of $25) and CGC's typically-lower PSA-vs-CGC sale-price ratio (modern Pokemon: ~0.7x; vintage: ~0.5x). The CGC equation often favors submission on smaller cards where PSA's math says no.

This is why I'm running 30% of my submissions through CGC now. The break-even threshold is lower.

— Jamie

The vault's grading calculator runs this exact math on every card you flag. The decision is yours; the math should be transparent.