The Pokémon repeat-sales index
Every other price on this site is a median across different copies of a card, so it moves when the mix moves. This one does not: it is built only from the same physical slab selling twice, matched on its PSA certification number. It is the narrowest measure we publish, and the only one a change of sample cannot fake.
How the index has moved · September 2017 to August 2026
Log scale: equal vertical distances are equal percentage moves. On a linear axis everything before 2024 would sit flat against the floor and the 2021 boom would disappear. The latest reading is the highest the index has been.
Returns by grade · the middle bucket, not a pooled average
Each tile is the median across that grade's (card, grade) buckets, not a median across sales: the dataset stores one figure per bucket rather than the underlying pairs, so a pooled median cannot be recovered. A mean would be worse — it would let an eight-pair bucket outvote a nine-hundred-pair one. The round-trip counts are there so you can weigh each row yourself; for today's price gap between grades, see the PSA 10 premium index.
Cards by realised return · 2,488 cards with 8+ round trips
Best median return
| Card | Median | Trips | Lost |
|---|---|---|---|
Celestial Storm · #108 | +1233% | 8 | 13% |
Team Rocket Returns · #40 | +1118% | 9 | 11% |
XY Black Star Promos · #XY119 | +1094% | 8 | 25% |
XY Black Star Promos · #XY35 | +1062% | 9 | 22% |
DP Black Star Promos · #DP27 | +1044% | 8 | 13% |
Dark Explorers · #54 | +748% | 10 | 30% |
Dark Explorers · #107 | +625% | 14 | 0% |
Delta Species · #63 | +472% | 10 | 0% |
Majestic Dawn · #100 | +447% | 10 | 20% |
Plasma Blast · #97 | +400% | 12 | 25% |
Unseen Forces · #55 | +379% | 12 | 8% |
Rising Rivals · #104 | +375% | 23 | 13% |
Worst median return
| Card | Median | Trips | Lost |
|---|---|---|---|
Base Set 2 · #44 | -50% | 11 | 55% |
Mega Evolution Black Star Promos · #38 | -47% | 8 | 100% |
Mega Evolution Black Star Promos · #37 | -42% | 8 | 100% |
Mega Evolution · #184 | -30% | 12 | 83% |
Surging Sparks · #245 | -23% | 29 | 72% |
Chilling Reign · #194 | -22% | 13 | 62% |
Prismatic Evolutions · #175 | -22% | 9 | 67% |
Rebel Clash · #203 | -19% | 15 | 60% |
Journey Together · #183 | -18% | 48 | 77% |
Guardians Rising · #22 | -17% | 15 | 67% |
Phantasmal Flames · #109 | -16% | 63 | 68% |
Phantasmal Flames · #126 | -16% | 16 | 81% |
A card needs at least 8 completed round trips to appear. Cards with three or four pairs produce far more spectacular numbers, and none of them mean anything — the dataset keeps them, this page does not print them. For raw sales volume by card, see the most traded Pokémon cards.
Own one of these? Track what your copies are worth, and what comparable slabs are actually selling for.
Questions
What is a repeat-sales index?
It measures price change using only the same object sold twice. Every other price figure on this site is a median across different copies of a card, so it moves whenever the mix of copies changes — the card did not change value, the sample did. A repeat pair is one physical slab, identified by its PSA certification number, at two points in time. It is the method behind the Case-Shiller house-price index, for the same reason. This index is built from 90,789 such round trips.
How much have graded Pokémon cards gone up?
On this index, September 2017 = 100 and August 2026 stands at 1,748 — about 17.5× over the period. The path was not smooth: the index peaked in the 2021 boom, gave most of it back through 2022 and 2023, and only passed its old high again later.
Do people make money reselling graded Pokémon cards?
The median round trip in this data returned +40.9% over a median hold of 170 days. But 24% of them lost money. Those two numbers belong together — a typical result of about +41% comes with roughly a one-in-four chance of selling for less than you paid, before selling fees.
Why might this overstate real returns?
Because a pair only exists if the slab sold again. Anyone holding a card at a loss, or who bought at the top and did not sell, never enters the data at all. This measures completed round trips — in effect, finished card flips — which is a real and useful thing to know, but it is not a forecast of what a holder should expect.
Where does the data come from?
PSA's Auction Price Report. A cert number is permanent, unique to one slab and grade-locked — regrading issues a new cert — so two sales under one cert are the same object at the same grade. Pairs held under 30 days are excluded, as are certs appearing under two different card ids, which are binding errors rather than resales.
Source: PSA Auction Price Report, 90,789 round trips across 4,032 cards. Minimum hold 30 days. Dataset built 24 August 2026. Index levels are estimated by a Bailey-Muth-Nourse regression over every pair at once; where a stretch of months is not connected by enough pairs to identify a level, no level is published for it rather than a guess. Returns are gross — they do not deduct seller fees, shipping or grading costs.