A graded slab is not the same market as a raw card. We measured how far apart they sit, and how much the gap moves from one card to the next.
Published 2 November 2026 by CompsAPI
Two things change when a card goes from raw to graded. The price moves up, and the spread of prices narrows. We measured both on real completed eBay sales over three years, on 2 November 2026.
Victor Wembanyama, 2023 Prizm base rookie. Raw means the bare search. Graded means the PSA 10 sales only, parsed from the title.
| Search | Sales | Median | 25th | 75th |
|---|---|---|---|---|
| raw query | 199 | $165.25 | $70.00 | $463.75 |
| PSA 10 only | 53 | $1,136.00 | $325.00 | $2,712.50 |
The PSA 10 median is $1,136.00, about seven times the raw median of $165.25. The raw search is wide, $70.00 to $463.75 at the quartiles, because it is every copy of the card at every grade and condition mixed together. Pulling out the PSA 10s isolates one grade, though the spread is still large, $325.00 to $2,712.50, which is a different lesson we come back to below.
Umbreon VMAX alternate art, the card people mean when they say moonbreon. Same three year window.
| Search | Sales | Median | 25th | 75th |
|---|---|---|---|---|
| raw query | 186 | $2,300.00 | $1,657.50 | $3,827.88 |
| PSA 10 only | 56 | $3,900.00 | $3,799.75 | $4,150.73 |
Here the PSA 10 premium is about 1.7 times, not seven. But look at the quartiles. The raw card runs $1,657.50 to $3,827.88, a wide band. The PSA 10 runs $3,799.75 to $4,150.73, a tight band around $3,900.00. Grading did not just raise the price, it collapsed the uncertainty. A sealed grade is close to a commodity, so its sales cluster.
This is the part that catches people out. The PSA 10 premium was about seven times on the Wembanyama and about 1.7 times on the Umbreon. You cannot take a raw price and multiply by a constant to get the graded price, or the other way round. The multiple depends on how hard the card is to grade clean, how many already sit in a slab, and how much the top grade is chased. It is a property of the specific card, measured per card, not a rule of thumb.
It follows that if you want a graded value you have to pull the graded sales, not estimate them. A raw median with a mental 3x on top will be badly wrong in both directions.
Ask for the card object and name the grade and grader. That parses the grade out of each title and returns only the sales that match.
curl -G https://api.compsapi.com/v1/comps \
-H "Authorization: Bearer YOUR_KEY" \
--data-urlencode "q=victor wembanyama 2023 prizm rookie" \
--data-urlencode "cards=1" \
--data-urlencode "grader=PSA" \
--data-urlencode "grade=10" \
--data-urlencode "days=1095"One caution from the Wembanyama numbers. Even inside PSA 10 the spread was $325.00 to $2,712.50, because titles are seller written and a base Prizm, a Silver Prizm and a numbered parallel can all read as prizm rookie. A clean grade filter narrows the field but does not finish the job, so quote the median and the spread together, and add the parallel or the card number when the band stays wide. We walk through that matching problem on cards here.
100 free requests a month · every filter on · no card
Get a free key