The Complete Guide to TCG Card Grading in 2026

Card grading is a paid service where a third-party company authenticates a trading card (confirms it isn't counterfeit, altered, or trimmed), assigns a numeric condition score, and encapsulates it in a sealed plastic case called a slab. The slab carries a unique cert number that buyers can verify on the grader's online registry.

Three companies dominate the TCG market: PSA (Professional Sports Authenticator), BGS (Beckett Grading Services), and CGC (Certified Guaranty Company). All three grade Pokémon, Magic: The Gathering, Yu-Gi-Oh!, One Piece, Lorcana, and most modern TCGs. PSA holds the largest market share for Pokémon by a wide margin; BGS is preferred for high-end vintage sports and select TCG investors who care about subgrades; CGC has the fastest turnaround and the lowest entry-tier price.

A grader evaluates the card across four condition axes — centering, corners, edges, and surface — and rolls those into a single 1-10 grade. A PSA 10 Gem Mint modern Charizard can be worth 8-15x the raw card. A PSA 9 is typically worth 1.5-3x raw. A PSA 8 or below often barely covers the grading fee. The financial leverage of grading is concentrated almost entirely in the top two tiers.

All three majors use a 1-10 scale where 10 is the top grade, but the scales are not interchangeable. A PSA 10 is the most liquid; a BGS 9.5 with high subgrades often outsells a BGS 10 because BGS 10s are rare enough that the comp market is thin.

PSA uses whole numbers 1 through 10. PSA 10 Gem Mint requires four sharp corners, full original gloss, perfect or near-perfect centering (commonly cited as 55/45 or better front, 75/25 or better back), and no print defects visible at standard viewing. PSA 9 Mint allows one minor flaw. PSA 8 NM-MT allows two minor flaws. PSA 7 down to PSA 1 track increasing wear.

BGS uses the same 1-10 ladder but issues four subgrades (one per axis) plus a final composite. A BGS 9.5 Gem Mint is roughly equivalent to a PSA 10 in market terms. A BGS 10 Pristine requires three subgrades of 10 and one of at least 9.5. A BGS 10 Black Label — all four subgrades at 10 — is genuinely rare and commands a multi-fold premium.

CGC uses 1-10 with half-point increments. CGC 10 Pristine maps roughly to BGS 10 Pristine; CGC 10 Perfect is their equivalent of Black Label. CGC 9.5 Mint+ sits between PSA 9 and PSA 10 in market value, which is part of why cross-grade arbitrage exists.

The practical takeaway: a 9 and a 10 can look identical to the naked eye but differ in price by 5-10x. Centering is the most common reason a card that "looks like a 10" comes back a 9.

Grading fees scale with the declared value of the card and the turnaround tier you select. Higher declared value means longer insured handling and higher fees; faster turnaround costs a multiple of the base tier.

Hidden costs people forget: two-way shipping ($15-40 depending on insured value), return-shipping insurance (typically 1-3% of declared value), and submission supplies (sleeves, semi-rigid Card Savers, team bags). For a 10-card submission, $40-80 in shipping and supplies is normal. Add seller fees at sale (eBay 13.25% final-value fee, processing) and the total cost-to-cash is often $30-50 per card before profit even starts.

Card grading is worth it when the expected graded value minus all costs exceeds the raw resale value by enough margin to justify the wait and the downside risk. The decision is a probability-weighted expected-value calculation, not a vibe.

The formula: Expected Graded Value = P(10) × Price(10) + P(9) × Price(9) + P(8) × Price(8) + P(≤7) × Price(raw). Subtract grading fee + two-way shipping + return insurance + marketplace fees + opportunity cost of cash tied up for 2-4 months.

Worked example — modern Pokémon $80 raw with $400 PSA 10 comp: Estimate P(10) = 30%, P(9) = 50%, P(8) = 20%. PSA 10 nets ~$340 after eBay fees; PSA 9 nets ~$130; PSA 8 nets ~$70. Expected gross = (0.30 × 340) + (0.50 × 130) + (0.20 × 70) = $181. Subtract $25 grading + $20 shipping = $136 net. Raw sells for ~$68 net. Edge: ~$68 per card, but only if your P(10) estimate is accurate.

Worked example — modern $30 card with $90 PSA 10 comp: Even at P(10) = 25%, expected gross is ~$45; subtract $40 in fees → $5 net. Raw sells for ~$25 net. Lose money on average. Grade only if you enjoy the gamble.

Worked example — vintage 1999 Charizard, raw $400, PSA 9 = $2,500, PSA 10 = $18,000: Even at P(10) = 5%, P(9) = 30%, expected value is $1,650 vs $340 raw. Grade every time. The asymmetry is decisive.

Two common mistakes: overestimating P(10) (most submitters think they have a 10; most don't), and using the eBay BIN max price instead of the median of recent sold listings. Sold listings are the only honest comp.

Graders evaluate every card across four condition axes: centering, corners, edges, and surface. Each axis has its own deduction ladder, and the final grade is bounded by the weakest axis — the bucket-floor rule that drives most disappointing grades.

Centering is measured as the ratio of opposing borders. PSA 10 centering threshold is commonly cited as 55/45 or better on the front and 75/25 or better on the back. Off-center is the single most common reason cards drop a tier.

Cornersare scored under angled light and magnification. Graders look for whitening (fiber exposure where ink has worn), softening (rounding), and dents. A single whitened corner visible at arm's length typically caps the card at PSA 8.

Edges are scanned across all four sides for edge whitening, chips, and roughness. Modern Pokémon black-bordered cards and MTG black-bordered cards punish edge whitening severely because the contrast makes it obvious.

Surface is the deepest axis and where most surprise downgrades come from. Graders look for print lines, print dots, scratches (especially on holofoil), indentations, scuffing, fingerprints, clouding of the gloss, and any sign of alteration. An altered card grades 0 / Authentic Altered or is rejected.

The grading fee is sunk the moment you ship; preparation is the only lever you control. The goal is to present the card exactly as it left the pack, with no human-introduced wear stacked on top of factory condition.

Inspect first.Examine the card under direct light at an angle. Use a 10x loupe or your phone's macro mode. Note every defect. If corners are already soft or the surface is scratched, grading is probably not worth it.

Do not clean. Solvents, erasers, magic-erasers, and microfiber-with-pressure all count as alteration. A card that looks dirtier ungraded is worth more than one detected as altered.

Sleeve and Card-Saver. Insert the card into a penny sleeve with the opening up, then into a semi-rigid Card Saver I with the opening at the top of the sleeve and the bottom of the Card Saver. This orientation lets graders extract the card from both ends.

Package for shipping. Two layers of cardboard sandwich, bubble wrap, and a rigid mailer. Declare honest values, and buy shipping insurance for the full declared value. USPS Registered Mail is the standard for high-value submissions.

AI pre-grading is a pre-screen step: you photograph a card with your phone, an AI vision model and a rule-based scoring engine analyze the four condition axes, and you get back a predicted PSA tier and confidence band — typically in under a minute, for around $1 per scan. It is not a replacement for PSA, BGS, or CGC. Its job is to answer one question: should I pay $25 and wait 3 months to send this card in?

AuraGradeis one such pre-grading tool. We score on a deterministic rubric of 50+ rules across the four axes — the same photo gives the same score every time. Centering uses pure ratio math from the card's color geometry, so it works equally on Pokémon yellow, MTG black, and Yu-Gi-Oh dark borders. Corner, edge, and surface defects are identified by Claude Opus 4.7 Vision and converted to deduction points by a fixed rubric.

What "±1 PSA tier" means in practice: if AuraGrade predicts PSA 9, the actual grader call lands at PSA 8, 9, or 10 in roughly 70-85% of internal-test cases — exactly the band where the $25 submission-fee decision changes. The remaining 15-30% can land a full tier outside that range; borderline cases (BGS 8.5 vs 9, light edge whitening, holo wear under angled light) are where any AI grader loses precision.

This is an engineering estimate on the underlying methodology, not a measured certification. The two components behind it are different in nature: centering is pure pixel math from a rectified LAB-chroma measurement — exact within optical resolution, no probabilistic component; corner / edge / surface defectsare identified by Claude Opus 4.7 Vision with bbox-annotated evidence — these carry the probabilistic component. Phone-camera scans cannot reach Gem Mint regardless of detected defects because hairline scratches and gloss loss are not visible in diffuse phone optics. The combined tier estimate will re-baseline once the first 50 real PSA returns arrive. Use AI pre-grading the way you'd use a home blood-pressure cuff: cheap, fast, useful for triage, not a substitute for the lab.

There is no universal best grader; the right choice depends on what you're grading and why. PSA wins for liquidity and resale velocity, especially on Pokémon and modern sports. BGS wins for top-tier vintage and any card where subgrades materially shift the comp. CGC wins for turnaround and entry-tier price, and has the strongest Magic ecosystem.

A practical rule: if you don't have a specific reason to use BGS or CGC, default to PSA. The liquidity premium on PSA slabs offsets fee and turnaround disadvantages for the vast majority of TCG cards.

Most disappointing grades trace back to a handful of repeated failures. Recognizing them ahead of submission is worth more than any other piece of grading advice.

• Underestimating centering.People eyeball centering and assume it's close enough. Graders measure with rulers. A 60/40 front center caps you at PSA 9, full stop.
• Holo scratches you can't see straight-on. Tilt a holofoil V or ex card under direct light. Hairline scratches that are invisible from above will reflect at an angle and cost you a tier.
• Edge whitening on black-bordered cards. A whitened edge that's invisible on a Pokémon yellow card is glaring on an MTG black-bordered card.
• Sub-millimeter corner dents. A corner that feels sharp can still be soft under magnification.
• Print lines and print dots.Factory artifacts are not the submitter's fault but still deduct. Reverse holos and full-art cards are especially prone.
• Back-of-card defects. Submitters obsess over the front and forget the back. PSA grades both sides; a back miscut or back surface scratch can drop the grade independently.