Across 7,270 cards graded by TCGAI.PRO between December 20, 2025 and October 22, 2026, the median TCGAI Score was 9 and the average 8.28. One card in five (20.6%) scored a perfect 10, and 60% scored 9 or higher. Modern sets scored far higher than vintage ones, and among imperfect cards, edges and surface were the sub-grades that most often held the score down.
These are AI-generated estimates, not official grades. See the methodology for how each score is produced and the notes on this data below before drawing conclusions.
Key numbers
- 7,270 cards with a TCGAI Score (December 2025 to October 2026)
- 9 median score, 8.28 average
- 20.6% scored a perfect 10
- 28.9% scored 9.5 or higher (Gem Mint range)
- 60.0% scored 9 or higher
- 15.0% scored below 7
Score distribution
| TCGAI Score | Cards | Share |
|---|---|---|
| 10 | 1,495 | 20.6% |
| 9.5 | 604 | 8.3% |
| 9 | 2,262 | 31.1% |
| 8.5 | 158 | 2.2% |
| 8 | 986 | 13.6% |
| 7 | 677 | 9.3% |
| 6 | 405 | 5.6% |
| 5 | 285 | 3.9% |
| 4 to 4.5 | 219 | 3.0% |
| 3 and below | 179 | 2.5% |
The biggest single group is the 9: almost a third of all cards. The gap between "very good" and "perfect" is where most of the value sits, because the same card in a 9 and a 10 can sell for very different prices. That gap usually comes down to one soft corner, a chipped edge or a surface scratch.
Results by game
| Category | Cards | Average score |
|---|---|---|
| Pokémon | 2,550 | 8.23 |
| Sports cards (all) | 3,524 | 8.03 |
| Football | 1,302 | 8.78 |
| Baseball | 1,143 | 7.14 |
| Basketball | 1,015 | 8.07 |
| One Piece | 278 | 9.17 |
The One Piece row counts cards identified as One Piece; some One Piece cards were labelled "other" by the identification, so One Piece is undercounted here (its sets are counted in full in the next table). Yu-Gi-Oh!, Magic: The Gathering, Dragon Ball and Lorcana have fewer than 25 cards each in this data, too few for a meaningful average. Baseball's average is pulled down by its two largest sets, Topps (6.05) and Topps Baseball (5.55); see the next section.
Vintage versus modern: the sets with the most cards
| Set | Game | Cards | Average | 9.5 or higher |
|---|---|---|---|---|
| 151 (Scarlet & Violet) | Pokémon | 145 | 9.14 | 47.6% |
| Phantasmal Flames | Pokémon | 111 | 7.72 | 16.2% |
| Mega Evolution | Pokémon | 84 | 8.48 | 29.8% |
| Ascended Heroes | Pokémon | 82 | 8.41 | 23.2% |
| Base Set (1999) | Pokémon | 75 | 6.32 | 2.7% |
| Scarlet & Violet Black Star Promos | Pokémon | 62 | 9.05 | 22.6% |
| Scarlet & Violet | Pokémon | 60 | 9.01 | 38.3% |
| Crown Zenith | Pokémon | 57 | 9.09 | 36.8% |
| Topps | Baseball | 165 | 6.05 | 6.1% |
| Topps Baseball | Baseball | 88 | 5.55 | 0% |
| Donruss | Baseball | 59 | 7.58 | 5.1% |
| Topps Chrome | Basketball | 53 | 8.97 | 49.1% |
| Awakening of the New Era | One Piece | 70 | 9.51 | 70.0% |
Sets with at least 40 graded cards. Pokémon Base Set, printed in 1999, averaged 6.32 and only 2.7% reached the Gem Mint range, against 9.14 and 47.6% for the modern 151 set. The most likely reason is simple: cards that have been handled, stored and traded for decades carry more wear.
Which sub-grade costs the most
Detailed sub-grades are stored for 2,472 of the cards (April to October 2026). For each card we take the lower of the front and back value in each category.
| Sub-grade | Average | Lowest (or tied lowest) on imperfect cards |
|---|---|---|
| Centering | 9.06 | 54% |
| Corners | 9.18 | 76% |
| Edges | 9.24 | 85% |
| Surface | 9.35 | 85% |
Of those 2,472 cards, 2,290 had at least one sub-grade below 10. Shares add up to more than 100% because several sub-grades often tie for the lowest value. Centering has the lowest average, while edges and surface are the ones most often at or tied for the bottom. In practice: check the borders first, because centering cannot be improved, then look hard at edges and surface under angled light.
What this means for collectors
- Plan for a 9, not a 10. Only one card in five scored a 10 here, and the sample probably leans toward cards their owners expected to grade well.
- Vintage needs a closer look. Pokémon Base Set scored almost three points lower on average than the modern 151 set. Pre-screening matters most where the expected grade is low compared with the grading fee.
- Edges and surface decide most 9-versus-10 outcomes. They are also the hardest to judge by eye, which is why a zoomed inspection pays off.
About this data
- Source: every card graded on TCGAI.PRO (web and Android app) from December 20, 2025 to October 22, 2026, excluding test gradings. One card had no score and is left out. Figures were extracted on October 23, 2026.
- Two versions of the method: cards graded before October 21, 2026 went through the previous version of our grading pipeline; later cards use the current one described in the methodology. Both score the same four sub-grades on the same scale, but they are not identical.
- Self-selected sample: collectors choose what to scan, usually cards they hope will grade well. These numbers describe the cards people pre-grade, not the hobby as a whole.
- Game and set labels come from the card identification. A few cards have no game or set and are excluded from those tables; set spellings that differ only in naming were merged.
- Not official grades: the TCGAI Score is an AI estimate. Third-party grades may differ. TCGAI.PRO is not affiliated with any grading company.
Grade your own cards free and see where they land before you pay to submit.