TAG Grading Centering: Pristine 10 vs Gem Mint 10, Sports vs TCG
TAG centering tolerances for every grade: ~51/49 for Pristine 10, ~55/45 for Gem Mint 10, the sports vs TCG back split, the 1000-point score, how TAG measures.
TAG Grading publishes the most granular centering tolerances of any major grader, and they are tighter at the top than most collectors expect. A TAG Pristine 10 (TAG Score 990 to 1000) requires the front image centered within about 51/49 and the back within about 54.5/45.5 for sports or 52/48 for TCG. A TAG Gem Mint 10 (950 to 989) allows about 55/45 on the front and 70/30 (sports) or 65/35 (TCG) on the back. A TAG Mint 9 (900 to 949) allows about 60/40 front and 90/10 (sports) or 75/25 (TCG) back.
Two things make TAG different from the PSA-style ladder you may already know. First, there is no 9.5: TAG's position is that a card better than Mint is Gem Mint, and a card better than Gem Mint is Pristine, with the 1000-point TAG Score doing the fine sorting. Second, TAG measures centering by machine. Its rubric and help articles describe machine-learning image capture, eight subgrades on a DIG report (front and back centering among them), and a rule that a Pristine front "generally cannot exceed a single measurement of 51/49." That word "single" tells you the worst axis governs, exactly as it does at PSA, but with the measurement done by software and then reviewed by a human before the grade is finalized.
Below is the full rubric in one table, the sports-versus-TCG back split explained, the arithmetic in millimeters, and what a deterministic measurement means for a card that sits right on a line.
The full TAG centering rubric by grade
All tolerances are from TAG's published rubric; TAG writes them with a tilde (~) to signal "approximately." Front tolerances apply to both sports and TCG. Where the rubric states a back figure, it is listed; below 8 the rubric states only the front.
| TAG grade | TAG Score band | Front centering | Back centering (sports) | Back centering (TCG) |
|---|---|---|---|---|
| 10 Pristine | 990–1000 | ~51/49 | ~54.5/45.5 | ~52/48 |
| 10 Gem Mint | 950–989 | ~55/45 | ~70/30 | ~65/35 |
| 9 Mint | 900–949 | ~60/40 | ~90/10 | ~75/25 |
| 8.5 NM-MT+ | 850–899 | ~62.5/37.5 | ~95/5 | ~85/15 |
| 8 NM-MT | 800–849 | ~65/35 | not stated | ~95/5 |
| 7.5 NM+ | 750–799 | ~67.5/32.5 | – | – |
| 7 NM | 700–749 | ~70/30 | – | – |
| 6.5 EX-MT+ | 650–699 | ~72.5/27.5 | – | – |
| 6 EX-MT | 600–649 | ~75/25 | – | – |
| 5.5 EX+ | 550–599 | ~77.5/22.5 | – | – |
| 5 EX | 500–549 | ~80/20 | – | – |
| 4.5 VG-EX+ | 450–499 | ~82.5/17.5 | – | – |
| 4 VG-EX | 400–449 | ~85/15 | – | – |
| 3.5 VG+ | 350–399 | ~87.5/12.5 | – | – |
| 3 VG | 300–349 | ~90/10 | – | – |
| 2.5 GOOD+ | 250–299 | ~92.5/7.5 | – | – |
| 2 GOOD | 200–249 | ~95/5 | – | – |
| 1.5 FAIR | 150–199 | ~98.33/1.67 | – | – |
| 1 POOR | 100–149 | may show a tiny sliver of border, but may not be miscut | – | – |
Notice the structure. From 8 downward, every half grade is a 2.5-point step on the front ratio: 65/35, 67.5/32.5, 70/30, 72.5/27.5, and so on. That regularity is a product of the numeric scoring system; it is much finer than the PSA ladder, where 7 is 70/30 and the next step down, 6, jumps to 80/20. It also means a card TAG scores as 7.5 (67.5/32.5 front) has nowhere to land on PSA's published ladder except "between 7 and 8."
For a side-by-side of all graders at every grade, see our card centering tolerances by grade master table.
Pristine 10 vs Gem Mint 10: the 51/49 line
The gap between TAG's two 10s is where most of the questions come from. Both say "10" on the label, but the centering requirements are very different.
| TAG Pristine 10 | TAG Gem Mint 10 | |
|---|---|---|
| TAG Score | 990–1000 | 950–989 |
| Front | ~51/49 | ~55/45 |
| Back, sports | ~54.5/45.5 | ~70/30 |
| Back, TCG | ~52/48 | ~65/35 |
TAG's help center is explicit about how strictly the front is held: "Front Centering generally cannot exceed a single measurement of 51/49." In millimeters, on a card with 5.0 mm of combined left-plus-right border (typical of an English Pokémon-style yellow border at roughly 2.5 mm per side), 51/49 means 2.55 mm vs 2.45 mm. That is a 0.10 mm difference between opposite borders. On a sports card with about 4.0 mm of combined border, it is 2.04 vs 1.96 mm, a 0.08 mm difference. Those are tolerances you cannot reliably confirm with a ruler; a digital caliper reading to 0.01 mm, or pixel counting on a high-resolution scan, is the minimum at home.
For comparison, 55/45 on the same 5.0 mm total is 2.75 vs 2.25 mm, a 0.50 mm difference, five times the Pristine allowance. So a card that looks "dead centered" to the eye, say 53/47, is a comfortable Gem Mint 10 at TAG and a failed Pristine.
Pristine also requires the overall score to reach 990, which means corners, edges and surface must be essentially perfect too. Centering within 51/49 is necessary, not sufficient. The reverse is also true: a card with 50/50 centering and one tiny corner touch can fall to 989 or below and receive Gem Mint 10.
Why the back tolerance differs for sports and TCG
TAG is the only major grader whose published rubric splits back centering by card type, and TCG backs are held tighter at every published level: 52/48 vs 54.5/45.5 for Pristine, 65/35 vs 70/30 for Gem Mint, 75/25 vs 90/10 for Mint 9, 85/15 vs 95/5 for 8.5, and a 95/5 back figure stated for TCG at 8.
TAG does not publish a rationale on the rubric page, so the reasoning below is process knowledge rather than an official statement. Most TCG backs, Pokémon's blue back being the obvious example, have a uniform printed frame with a consistent border all the way around, so a back measurement is clean and repeatable, and collectors actively look at backs. Many sports backs are text-heavy designs where the printed area is less regular, and hobby convention has always treated sports backs more loosely; PSA, for instance, allows a 75/25 back on a 10 and 90/10 on everything from 9 down to 4. TAG's sports back numbers at Gem Mint (70/30) and Mint (90/10) sit close to that convention, while its TCG back numbers are a notch tighter.
The practical consequence: if you are choosing between submitting a TCG card to TAG or elsewhere, measure the back. A Pokémon card with a 70/30 back fails TAG Gem Mint (65/35 limit) but passes PSA's Gem Mint back limit of 75/25. Our guide to back centering tolerances across graders walks through more of these cases.
Worked examples in millimeters
Use the standard formula: measure opposite borders, add them, divide the larger by the total. Report the worse axis.
Example 1: modern TCG card, front. Left 2.35 mm, right 2.65 mm. Total 5.00 mm. 2.65 / 5.00 = 53%. Left-right is 53/47. Top 2.20 mm, bottom 2.80 mm. 2.80 / 5.00 = 56%. Top-bottom is 56/44. Worst axis 56/44: fails Pristine (51/49) and, read strictly, sits just outside Gem Mint (55/45). At TAG the rubric's "" leaves the grader some room, but this is exactly the kind of card where the machine measurement decides and you should not expect leniency. Expected centering outcome: Mint 9 territory on centering, unless the measurement lands on the 55 side.
Example 2: same card, back (TCG). Left 1.70 mm, right 3.30 mm. Total 5.00 mm. 3.30 / 5.00 = 66%. Back is 66/34. TCG Gem Mint back limit is ~65/35, so this back fails Gem Mint by a point and is a Mint 9 back (limit 75/25). For a sports card with the same measurement, 66/34 passes the 70/30 Gem Mint sports limit.
Example 3: sports card, Pristine attempt. Left 2.02 mm, right 1.98 mm. Total 4.00 mm. 2.02 / 4.00 = 50.5%. Left-right 50.5/49.5, within 51/49. Top 2.30 mm, bottom 2.20 mm. Total 4.50 mm. 2.30 / 4.50 = 51.1%. Top-bottom 51.1/48.9. That single measurement exceeds 51/49 by a tenth of a point. By TAG's wording ("generally cannot exceed a single measurement of 51/49") this card is at real risk of Gem Mint rather than Pristine on centering alone, even though the other axis is nearly perfect. This is the worst-axis rule at work, and it is why collectors report TAG Pristine as hard to hit.
Example 4: the 8.5 step. Front 2.65 vs 1.35 mm. Total 4.00 mm. 2.65 / 4.00 = 66.25%. That is 66.25/33.75: past the 8.5 limit of 62.5/37.5 and past the 8 limit of 65/35, so centering supports a 7.5 (67.5/32.5 limit). At PSA, 66/34 is between the 8 line (65/35) and the 7 line (70/30) and would read as a 7 on centering, or an 8 under PSA's 5% leeway for eye appeal. TAG's finer steps give you the half grade.
You can get these numbers from a straight-down photo with the free card centering calculator, which reports left/right and top/bottom ratios for front and back and maps them to the TAG bands as well as PSA, BGS, CGC and SGC. For Pristine-level decisions, though, treat any app or ruler reading within a point of 51/49 as "too close to call" and use a caliper or scan.
How TAG measures: machine capture, human finalization
TAG describes its process in several official help articles, and the combination matters for how you should think about borderline cards.
- Capture. TAG uses "machine-learning image capture software which analyzes cards for defects." Its DIG page describes "Photometric Stereoscopic Imaging" and a viewer that zooms "up to 800x magnification."
- Scoring. Cards receive a TAG Score "on a 1000-point scale, and an industry 10-point equivalent." The DIG report lists "eight subgrades: front and back centering, front and back corners, front and back surface, and front and back edges."
- DINGS. TAG's term for flagged defects is DINGS, "Defects Identified of Notable Grade Significance," each located on the card image in the report.
- Human review. "All defects identified by TAG's grading system are thoroughly reviewed by human grading experts before Grade Finalization."
What this means for centering specifically: the ratio on your DIG report is a measured value, not an eyeballed one, and it is reported separately for the front and the back. If your card measures 56/44 on one axis, that number will be on the report, and the Gem Mint line is 55/45. Where PSA's standards page builds in an explicit eye-appeal variance clause and a 5% leeway for PSA 7 and better, TAG's published rubric offers only the "~" on each figure. Collectors report TAG as the most consistent of the major graders on centering for exactly this reason, and also the least forgiving of a card that is a point over the line. We compare that to PSA's human-judgment approach in how does PSA actually measure centering?.
One nuance worth repeating: the 1000-point score means a Gem Mint 10 at 952 and a Gem Mint 10 at 988 are both "10" on the 10-point equivalent, but they are not the same card, and buyers who read DIG reports know it. A centering subgrade that is merely inside the 55/45 line will cost points relative to a 50/50 card even within the same letter grade.
Why there is no 9.5 and what that does to your expectations
TAG's help center answers this directly: "TAG believes cards which exceed our standard for a Mint condition card are Gem Mint cards, and cards which exceed our criteria for Gem Mint are Pristine cards." The 1000-point score replaces the half grade: a 949 and a 901 are both Mint 9 on the label, but the report tells them apart.
For centering, this removes a familiar buffer. At BGS, a 55/45 front is a 9.5 Gem Mint and a 50/50 front is needed for a 10; collectors with a 53/47 card know they are chasing a 9.5. At TAG, a 53/47 front is a Gem Mint 10 (if the rest of the card scores 950 or better), and the only thing above it is the 51/49 Pristine tier. So a very well centered card that would be "just a 9.5" elsewhere can be a 10 at TAG, while a 56/44 card that might squeak through on eye appeal elsewhere has no 9.5 to fall back on and goes to Mint 9. The centering-only equivalence between a TAG 10, a PSA 10, a BGS 9.5 and a CGC 10 is laid out in our centering crosswalk.
TAG vs PSA on centering, side by side
| TAG Pristine 10 | TAG Gem Mint 10 | PSA 10 Gem Mint | TAG Mint 9 | PSA 9 Mint | |
|---|---|---|---|---|---|
| Front | ~51/49 | ~55/45 | ~55/45 | ~60/40 | ~60/40 |
| Back, sports | ~54.5/45.5 | ~70/30 | 75/25 | ~90/10 | 90/10 |
| Back, TCG | ~52/48 | ~65/35 | 75/25 | ~75/25 | 90/10 |
| Eye-appeal variance | "~" only | "~" only | Explicit footnote: small variance at grader's discretion | "~" only | Explicit 5% leeway for PSA 7+ |
| Half grade above | – | Pristine tier instead | none (PSA half points run 2 to 9) | none; score band instead | none |
Three takeaways. The front line for a 10 is the same 55/45 at both companies. TAG is stricter on the back, especially for TCG (65/35 vs 75/25 for the 10; 75/25 vs 90/10 for the 9). And TAG's Pristine tier has no PSA analogue at all, so a 50/50 card has a label upside at TAG that PSA cannot offer.
Context on volume, since people ask whether TAG slabs are common enough to matter: GemRate's 2025 totals (reported by cllct/Yahoo) put TAG at about 440,000 cards graded, up 83% year over year, against PSA's 19.26 million, and GemRate's July 2026 recap listed a record 57,000 cards for TAG in that month. TAG is small next to PSA but growing quickly, and it describes itself as founded in 2013 and based in Los Angeles.
Submitting to TAG with centering in mind
- Measure both axes, front and back, before you submit. Use a caliper or a scan for anything you hope is Pristine; an app or ruler is fine for separating 9 from 10 candidates.
- Know which back rule applies. TCG cards face 65/35 (Gem) and 52/48 (Pristine) on the back; sports cards face 70/30 and 54.5/45.5.
- Treat the worse axis as your number. A 50/50 left-right does not rescue a 56/44 top-bottom.
- Do not count on leeway. The rubric uses "~", but the measurement is machine-driven and collectors report TAG as the least forgiving grader at the line.
- Pick the tier that gives you the data you want. As of August 2026, TAG's help center lists tiers from Basic at $22 (45+ days, minimum 10 cards) through Priority at $149 (5 days, and TAG states this tier includes the 1000-point score and DIG+ report) and Walkthrough at $299, though TAG's pricing page currently shows regular tiers paused with limited Priority grading available. Check the official page before submitting.
- Read the DIG report when the card comes back. The front and back centering subgrades tell you exactly where the card landed relative to 51/49 and 55/45, which is more than any other grader's label gives you.
Bottom line
- TAG Pristine 10 (990–1000) needs about 51/49 front and 54.5/45.5 (sports) or 52/48 (TCG) back; Gem Mint 10 (950–989) needs about 55/45 front and 70/30 (sports) or 65/35 (TCG) back; Mint 9 needs about 60/40 front and 90/10 or 75/25 back.
- Front tolerances are the same for sports and TCG; only the back differs, and TCG backs are always tighter.
- The worst single measurement governs, and below 8 the front ladder steps down in 2.5-point increments every half grade.
- There is no 9.5; the 1000-point TAG Score does the fine sorting, and the DIG report shows front and back centering as two of eight subgrades.
- Centering is measured by machine-learning imaging and finalized by human graders, so expect consistency and little leeway at the line.
- On the front, TAG Gem Mint and PSA 10 share the 55/45 line; TAG is stricter on the back and offers a Pristine tier PSA does not have.
Frequently asked questions
What centering does TAG require for a Pristine 10?
TAG's rubric requires the front image centered within approximately 51/49 and the back within approximately 54.5/45.5 for sports or 52/48 for TCG cards. TAG's help center adds that front centering generally cannot exceed a single measurement of 51/49, meaning each axis is checked separately and the worst one governs. A Pristine 10 also needs a TAG Score of 990 to 1000 out of 1000, so centering is necessary but not sufficient.
What centering does TAG require for a Gem Mint 10?
TAG Gem Mint 10 covers scores of 950 to 989 and requires front centering within approximately 55/45, with the back within approximately 70/30 for sports cards or 65/35 for TCG cards. The 55/45 front figure matches PSA's current Gem Mint 10 standard, but TAG's back requirement is tighter than PSA's 75/25, and TAG splits the back tolerance by card type where PSA does not.
Why does TAG not have a 9.5 grade?
TAG's help center explains that TAG believes cards which exceed its standard for Mint are Gem Mint cards, and cards which exceed its criteria for Gem Mint are Pristine cards. Instead of a 9.5 tier, TAG publishes a 1000-point TAG Score, so two Mint 9 cards can be told apart by score (for example 905 versus 945) without inventing a half grade. The 10-point equivalent is there for comparison with other companies.
How does TAG measure centering?
TAG uses machine-learning image capture software that analyzes cards for defects, with what it calls Photometric Stereoscopic Imaging and zoom up to 800x magnification. Centering is reported as front and back subgrades on the DIG report alongside corners, edges and surface, eight subgrades in total. TAG states that all defects identified by its system are reviewed by human grading experts before the grade is finalized, so the measurement is machine-driven but the final call has human oversight.
Does TAG treat sports cards and TCG cards differently for centering?
Only on the back. Front tolerances are the same for both: about 51/49 for Pristine, 55/45 for Gem Mint, 60/40 for Mint 9, and so on down the ladder. Back tolerances differ: Pristine is about 54.5/45.5 for sports versus 52/48 for TCG, Gem Mint is 70/30 versus 65/35, Mint 9 is 90/10 versus 75/25, and 8.5 is 95/5 versus 85/15. TCG backs are held to a tighter standard at every published level.
Is a TAG 10 the same as a PSA 10 for centering?
A TAG Gem Mint 10 uses the same approximate 55/45 front tolerance as PSA's Gem Mint 10, so on the front they line up. On the back TAG is stricter, 70/30 for sports and 65/35 for TCG versus PSA's 75/25. TAG also has a Pristine 10 tier above Gem Mint with a 51/49 front requirement, which has no PSA equivalent because PSA's scale stops at 10 with no pristine tier.
Sources & further reading
- TAG Grading Rubric (official 1000-point bands and centering tolerances)
- TAG Help Center – How do I get a Pristine 10 grade?
- TAG Help Center – Why is there no 9.5 on the TAG grading scale?
- TAG Help Center – What is the TAG DIG report?
- TAG Help Center – Is TAG Grading AI?
- PSA Grading Standards (for comparison)
