Sources: Beckett, CGC, TAG. Daily snapshots, diffed. PSA per-card data is not published here. Their population report sits behind a login, so PSA appears only in rates and comparisons, never as a card list.
Your odds of a ten depend more on which grader you pick than on which card you send.
29,792,188
graded cards tracked, four graders
4.9x
Beckett against CGC, same cards
1.1x
CGC against TAG, same cards
Four graders, four scales
Each grader is scored against its own top grades here, on whatever cards it happens to receive. That makes this a description of four scales rather than a ranking, and the next section is what happens when you hold the cards still.
Grader | Top-grade rate | Top grade | Graded |
|---|---|---|---|
PSA | 64.9% | 10,931,548 | 16,852,923 |
TAG | 58.6% | 335,179 | 571,899 |
CGC | 47.8% | 5,699,335 | 11,923,298 |
Beckett | 28.3% | 125,744 | 444,068 |
Each grader is measured against its own top grades, not a shared scale. CGC counts Perfect 10, Pristine 10 and Gem Mint 10; TAG counts 10 and 10 Pristine; Beckett counts Pristine 10 and Black Label. Treat this as a description of each grader's own scale rather than a ranking: Beckett looks far stricter here only because a BGS 9.5, which the market treats as the equivalent tier, is not counted. Count it and Beckett moves ahead of CGC. These figures also cover different mixes of cards, which moves them again.
Same cards, different graders
Beckett returns its top grade on 10.2% of the very cards CGC returns its own top grade on 50.6% of. Same physical cards, 4.9 times the odds, measured across 1,932 of them.
At the other end, CGC and TAG land within a few points of each other, 65.2% against 68.7%. Two graders, effectively one standard.
Pair | Top-grade rate | Gap | Shared cards |
|---|---|---|---|
CGC vs PSA | 68.7% / 79.8% | 1.2x | 3,926 |
Beckett vs CGC | 10.2% / 50.6% | 4.9x | 1,932 |
CGC vs TAG | 65.2% / 68.7% | 1.1x | 1,821 |
Beckett vs PSA | 11.9% / 44.3% | 3.7x | 1,749 |
PSA vs TAG | 82.0% / 74.2% | 1.1x | 1,337 |
Beckett vs TAG | 9.4% / 45.4% | 4.8x | 733 |
This table and the corpus rates above it disagree, and the gap is the point. Beckett sits at 28.3% across everything it grades but 10.5% on cards someone else also graded. The corpus figure flatters it, purely because no two graders receive the same pile of cards.
Read across a row, never down a column: PSA alone runs from 44.3% to 82.0% depending on which grader it is being measured against, a 38 point swing on the same underlying standard.
Each pair is measured on its own overlap, so a grader's figure changes from row to row. That is the point: the only honest comparison is within a row, never down a column.
One set ate the week
All 5 of these cards are CGC, and that is a fact about our week rather than about the market. CGC was the only grader with a clean week. TAG had 2 scrapes that did not finish and 3 flagged by the health guard, Beckett had 4 flagged by the health guard. Read this as what CGC graded, not as what the hobby graded.
23,931 of the 45,489 gradings in this list came from one set, Mega Evolution / Mega Dream ex - M2a - Japanese. That is bulk submission, not broad demand. At most 2 cards from any single set are shown so the rest of the market stays visible.
Card | Added | Pop now |
|---|---|---|
199/193 Psyduck | +2,659 | 16,890 |
231/193 Mega Scrafty ex | +1,686 | 10,735 |
013/080 Mega Charizard X ex | +1,104 | 38,084 |
043 Rowlet | +740 | 6,568 |
049/080 Gengar | +656 | 5,508 |
Beckett also added 333 gradings across 1 set that it has not attributed to a named card. Real population, but not tied to a specific card, so it is not in the table.
Read from the top 1,000 changes per day, so the table above is the head of the list rather than a census. More cards moved than are shown.
Where the rare stuff lands
All 3 of these first grades are Beckett, and that is a fact about our week rather than about the market. TAG had 2 scrapes that did not finish and 3 flagged by the health guard. Read this as what Beckett graded, not as what the hobby graded.
3 of 3 are non-English printings (German, Italian, Spanish), which is the corner of the market almost nobody grades at volume.
Card | Grade | Pop |
|---|---|---|
GG33 Poochyena GGUR | 9 | 1 |
125 Mega Gengar ex RR | 8.5 | 1 |
046 Chikorita STAMP [30th Anniversary]/(First Partner Illustration Collection - Series 2) | 9.5 | 1 |
Filtered to cards appearing with a population of 3 or fewer. A card appearing with hundreds already graded is one we are seeing for the first time, not one the world graded for the first time.
New set pages
Set | Cards |
|---|---|
Promotional and Prize / Premium Coll. Live Action Edition Vol.2 - Japanese | 14 |
Scarlet & Violet / Eeveelution Coin Set - Simplified Chinese | 9 |
2026 Pokemon Card Game MEGA Storm Emeralda Japanese | 5 |
2011 Data Carddass One Piece Berry Match IC! Wave 2 Japanese | 3 |
39 new set pages in total; the 4 with the most cards are listed.
What we are building next
We started snapshotting on 13 August 2026, so everything above is a census rather than a trend. Population reports are cumulative and nobody publishes yesterday's, which means a day not captured is gone for good. That is what we have been fixing first.
Scarcity velocity, from mid October. Which top-grade populations are growing fastest, so you can see the scarcity that is evaporating rather than the scarcity that already has.
Crossover detection, from late November. Cracked and resubmitted slabs, spotted by watching one grader's population stall while another's accelerates on the same card. Tracking four graders at once is what makes this computable at all.
More of the catalogue, continuously. Just under half the corpus is currently matched to a canonical card. Every point we add widens every same-card comparison in this issue.
Methodology
Window: 2026-08-17 to 2026-08-23.
17 scrapes ran in this window: 8 clean, 7 flagged by the health guard, 2 that did not finish. CGC 5/5 clean, TAG 2/7 clean, Beckett 1/5 clean.
Not every grader reported on every day of this window, so day-level totals are not comparable across graders.
2 scrapes did not finish (TAG 2026-08-18, TAG 2026-08-21), so those days are not counted anywhere in this issue and the week's total is not shown.
Beckett caps each set page at 50 card rows and its own totals are computed over only those rows, so we cannot detect the cut from the data. 99 Beckett sets are affected. Beckett population here is a floor, not a total, and no Beckett market share is calculated.
Population reports are cumulative. Everything here is a difference between daily snapshots, not a figure any grader publishes directly.
The Graded Report, issue 1, 23 Aug 2026. Population data from CGC, TAG, Beckett and PSA, snapshotted and diffed. Not affiliated with PSA, CGC, TAG, Beckett, Pokémon or The Pokémon Company. Nobody can buy a mention here.