Check both sides for PSA 10
Load the 53/47 front and 71/29 back specimen, see which side limits the grade, and compare a PSA submission with selling raw using a break-even table for raw prices from $10 to $500.
Open the PSA 10 calculatorFree tool: photos stay in this browser
Flatten a phone photo, drag eight border lines under a 400% loupe, and read live L/R and T/B centering with PSA, BGS, CGC and SGC ceilings for front and back.
Specimens
400 % wells4 mm per corner
amber = whitening
Drag a line or press the arrow keys on it: 1 px, Shift 10 px, Alt 0.1 px.
Correction check · L/R
Grade lamps · front + back combined
PSA 10 limit:
Back not measured — lamps use the front only. Add the back to merge it.
free Proof of this card
One side, banner-stamped ·
PRO Full-resolution files
The free proof stops at 1000 px with a PRE-GRADE PREVIEW banner. Pro redraws these exact lines as a 2000 px, 300 DPI listing image, a front-and-back certificate PDF and a 30-card ROI workbook — from $4.99.
Your card
Specimen loaded: 51/49 Gem Mint 10 Holo Specimen.
Centering estimate only, not a grade. Not affiliated with PSA, Beckett (BGS), CGC, SGC, The Pokémon Company, Topps or Panini.
Short answer
For a PSA 10, the front must measure 55/45 or better on both axes and the back 75/25 or better. Measure the four borders, divide each one by the sum of its pair, and compare the larger share. This Pokemon card centering tool does that from one photo: it flattens the shot, snaps eight lines and lights the PSA, BGS, CGC and SGC lamps.
The built-in Gem Mint specimen reads 51.2/48.8, which clears PSA 10, CGC 10 and SGC 10 but stops at BGS 9.5, because BGS 10 and CGC Pristine 10 need 50/50. Every grade down to 5 is on the centering standards chart, and cards without a printed border are covered in the full art centering guide.
How it works
Drop a JPG, PNG or WebP up to 15 MB. Four corner handles snap to the card; adjust them and the 4-point warp maps it onto a true 63.5 × 88.9 mm plane.
Outer lines sit on the cut edge, inner lines on the printed frame. Drag them under the 400% loupe or nudge with arrow keys: 1 px, Shift 10 px, Alt 0.1 px.
L/R and T/B update as you drag. Six lamps show pass, amber borderline or fail for PSA 10, BGS 10, BGS 9.5, CGC Pristine 10, CGC 10 and SGC 10.
Perspective
Tilt the phone and the near side of the card covers more pixels than the far side, so its border reads wider. On the built-in tilted specimen, a card that truly measures 52.0/48.0 reads 61.0/39.0 when you measure straight across the raw photo. That nine-point drift drops its ceiling from PSA 10 to PSA 8, past even the 60/40 limit for a 9.
The 4-point homography fixes this by mapping the four card corners onto a flat rectangle at 20 px per mm, so every millimeter of border gets the same number of pixels. Toggle perspective correction off and on: the raw and corrected readings stay side by side, and the Net EV on that specimen swings from −$4.00 to +$44.00.
Corner handles find the card outline automatically. On rounded corners they are fitted from the four straight edges, so each handle lands where two edges would meet. If the lines still look skewed after the warp, the rotation input trims ±5.0° in 0.1° steps. Then the card centering calculator measures a true rectangle, not a trapezoid.
Front and back
PSA 10, BGS 9.5, CGC 10 and SGC 10 all allow 55/45 on the front, so one front reading answers four questions. The backs differ: PSA and CGC accept 75/25, Beckett's current page gives 60/40 for BGS 9.5, and SGC publishes one figure per grade without naming a side, so the table shows its back as not stated.
The strict tiers sit above them. BGS 10, the centering a Black Label needs, takes 50/50 on the front and 55/45 on the back. CGC Pristine 10 takes 50/50 on the front and states no back figure, so that lamp is judged on the front alone. Every row lists its source page and the 2026-10-31 check date.
The differences matter on a lopsided back. The dual-side specimen, 53/47 front and 71/29 back, passes PSA 10 and CGC 10, yet 71/29 exceeds the 70/30 back Beckett lists for BGS 9, so its BGS centering ceiling drops to 8. The lamp matrix and this table read the same data file, so screen and chart always agree.
| Tier | Front | Back | What it takes |
|---|---|---|---|
| PSA 10 | 55/45 | 75/25 | PSA writes “approximately” |
| BGS 10 | 50/50 | 55/45 | Black Label also needs 10 on all four subgrades |
| BGS 9.5 | 55/45 | 60/40 | Both axes on the front |
| CGC Pristine 10 | 50/50 | — | Back figure not published |
| CGC 10 | 55/45 | 75/25 | Both axes on the front |
| SGC 10 | 55/45 | — | Side not specified by SGC |
Sources, checked 2026-10-31: PSA (www.psacard.com/gradingstandards); Beckett (BGS) (www.beckett.com/grading); CGC Cards (www.cgccards.com/card-grading/grading-scale/); SGC (gosgc.com/card-grading/scale). “—” means the company publishes no figure for that cell.
Expected value
Net EV = P10 × V10 + (1 − P10) × V9 − raw − fee. V10 and V9 are the PSA 10 and PSA 9 prices you enter, raw is what the card sells for ungraded, and fee is the per-card grading cost, $19 by default with $15 and $25 shortcuts. Fold shipping and insurance into the fee if you want them counted.
P10 is the chance of a 10: your baseline gem rate, 60% by default, times a front, a back and a corner factor. The front factor is 1 up to 52.5, 0.5 at 55.0 and 0 from 56.0. The back factor is 1 up to 70, 0.5 at 75 and 0 from 76. The baseline is your estimate for surface and edges.
Break-even gem rate = (raw + fee − V9) ÷ (V10 − V9). On the Gem Mint specimen, raw $59, PSA 10 $300, PSA 9 $100 and a $19 fee give P10 60.0% and Net EV +$142.00. Because a PSA 9 alone covers raw plus fee, break-even shows 0.0%. Every number updates live and stays free for single cards.
Loss traps
The classic trap is a card that looks centered at arm's length. The off-center specimen reads 58.1/41.9, which caps PSA at 9 and sets P10 to 0. With raw $44, PSA 9 $45 and a $19 fee, Net EV is −$18.00 and the call is SELL RAW. It would need a 13.3% gem rate just to break even.
The second trap is an uncorrected photo, which can flip the call either way: the tilted specimen reads −$4.00 raw and +$44.00 flattened. The third is skipping the back, since a 71/29 reverse trims the back factor to 0.9. The fourth is a cheap card: at $10 raw and a $19 fee, with PSA 9 at 1.2× raw and PSA 10 at 3×, break-even needs 94.4%.
Corners are the last trap. A whitened corner lowers P10 even when centering is perfect. The off-center specimen's bottom-right corner scores 3.3% on the Corner Whitening Index, which on its own would cut the corner factor to 0.54. Check all four loupes before you count on a 10, and lower the baseline for cards you have not inspected under light.
| Trap | Reading | Net EV | Call |
|---|---|---|---|
| Eyeballing a 58/42 front as “close enough” | 58.1 / 41.9 → PSA 9 cap | −$18.00 | SELL RAW |
| Measuring a tilted phone photo | 61.0 / 39.0 raw vs 52.0 / 48.0 flat | −$4.00 vs +$44.00 | SELL RAW vs SUBMIT TO PSA |
| Skipping the back | back 76.0 / 24.0 instead of 71.0 / 29.0 | −$9.00 vs +$73.62 | SELL RAW vs SUBMIT TO PSA |
| Higher fee, lower gem rate | $25 fee, 50 % baseline | +$53.85 vs +$73.62 | SUBMIT TO PSA |
| PSA 9 worth less than raw + fee | break-even gem rate 13.3% | −$18.00 | Only a 10 pays |
Net EV = P10 × PSA 10 price + (1 − P10) × PSA 9 price − raw price − grading fee, with P10 from the measured centering.
Corners
Each corner gets a 4 × 4 mm window on the flattened card. Inside it, only a 0.75 mm band along the cut edge is scored, with the 3 mm corner radius masked so the table behind the card stays out. A pixel counts as white when it is near gray and bright: at least 200 on a 0–255 scale and 45 above the band median.
CWI is white pixels divided by band pixels, shown to one decimal. Up to 1.0% keeps the corner factor at 1; each further point removes a fifth, so 2.0% gives 0.8, 3.3% gives 0.54 and 6.0% or more gives 0. The worst corner across front and back sets the factor, so one bad corner outweighs three clean ones.
White-bordered stock reads n/a, because a band whose median brightness is 215 or higher has no dark ink for whitening to show against. Those corners drop out of P10, which is why the vintage specimen shows n/a on all four. Inspect them by eye in the 400% loupes, and lower the baseline gem rate if you find wear.
| Worst corner CWI | Corner factor | P10 at 60 % baseline, perfect centering | Typical look |
|---|---|---|---|
| 0.5 % | 1.00 | 60.0% | Clean corner |
| 1.0 % | 1.00 | 60.0% | A fleck at most — no penalty |
| 2.0 % | 0.80 | 48.0% | Visible flecks under the loupe |
| 3.0 % | 0.60 | 36.0% | Clear whitening on one corner |
| 4.0 % | 0.40 | 24.0% | Heavy whitening |
| 6.0 % | 0.00 | 0.0% | Corner damage dominates |
CWI = whitening pixels ÷ band pixels in the 0.75 mm edge band of each 4 mm corner window. White-bordered cards show n/a and are not penalised.
Exports
Every single-card measurement in this card centering tool online is free with no sign-up: perspective correction, eight lines, the loupe, the lamp matrix, corner whitening and single-card EV. Free exports are a one-side proof PNG up to 1000 px wide, 1000 × 1464 px on a standard card, with a PRE-GRADE PREVIEW banner, plus a three-line TXT of L/R, T/B and the worst axis.
Pro adds three files. The batch ROI workbook queues up to 30 cards and exports a CSV with each card's ceilings, P10, Net EV, break-even and call, plus a PSA packing-slip PDF of the cards marked SUBMIT. The pre-grade certificate is a front-and-back PDF with mm and pixel deviation tables, plus a 300 DPI four-corner 400% loupe composite PNG.
The seller listing proof is a 2000 × 2000 px, 300 DPI PNG for eBay, TCGplayer or Whatnot with the gauge, a corner sidebar and a CENTERING nameplate. Pro costs $4.99 for a 7-day pass (was $9.99), $7.99 a month (was $15.99) or $19.99 lifetime (was $39.99). For scale, PreGradeCards charges $9.99 for 25 cards.
Card sizes
Pick the template before you flatten. Standard 63.5 × 88.9 mm, 2.5 × 3.5 in, fits Pokémon and most sports cards on a 1270 × 1778 px plane. Japanese size, 59 × 86 mm, fits the smaller stock some TCGs use. Vintage tall, 67 × 95 mm, fits older baseball cards and loads by default on the sports card calculator.
The template sets the rectangle the warp maps the corners onto, so it changes the millimeter readouts, not the ratio. A 64 px left border reads 3.20 mm on the standard plane and 2.97 mm on the Japanese one, while the L/R share stays put. The certificate and Manual mode both work in millimeters, so match the template to the card.
Where it fits
Load the 53/47 front and 71/29 back specimen, see which side limits the grade, and compare a PSA submission with selling raw using a break-even table for raw prices from $10 to $500.
Open the PSA 10 calculatorUse the vintage 67 × 95 mm template for tall cards with thick white borders, deal with sleeves and top loaders, and export a centering proof for eBay, TCGplayer or Whatnot listings.
Open the sports card calculatorCheck a card against 50/50 at 0.1% precision, read how many points it sits over, and see when the call switches from SUBMIT TO PSA to SUBMIT TO BGS.
Open the BGS calculatorSee PSA, BGS, CGC and SGC front and back limits side by side, with sources, check dates and the best grade a 51/49, 56/44 or 62/38 card reaches at each company.
View the standards chartFAQ
It is if the front measures 55/45 or better on both axes and the back 75/25 or better. Upload a photo, let the corner handles flatten it, and check the PSA 10 lamp: green passes, amber means less than one point over, and red means a PSA 9 or lower ceiling.
A card grader such as PSA, BGS, CGC or SGC handles the physical card and assigns one grade from centering, corners, edges and surface. A centering tool measures the border ratios from a photo and tells you the best grade centering allows, so you can decide which cards are worth the fee.
Yes, for the parts a photo can measure: centering on both sides, a corner whitening index and the money side of a submission. Surface scratches and edge chips still need your own eyes under good light, and the baseline gem rate input is where that judgment goes.
No. Photos are decoded and measured in your browser, and the free tool sends nothing. Only when an unlocked Pro user exports a certificate or listing image is the corrected card image sent to this site's server, without the original photo or EXIF data, processed in memory and not stored.
Accurate enough for the 55/45 line once the photo is flattened. The gauge works on a 20 px per mm plane in 0.1 px steps, but real accuracy depends on how many photo pixels cover the card. Fill the frame, shoot square to the card and compare the raw and corrected readings.
No account and no limits for single cards. Upload, flatten, measure both sides, read all six lamps, check corners and get the Net EV for free. Sign-in only matters for the Pro exports: the 30-card workbook, the pre-grade certificate and the 2000 px listing proof.
Three templates: standard 63.5 × 88.9 mm for Pokémon and most sports cards, Japanese size 59 × 86 mm for smaller TCG stock, and vintage tall 67 × 95 mm for older baseball cards. Switching templates changes the millimeter readouts while the ratios stay the same.
Amber is borderline: the card fails that tier but sits less than one point over its limit on each measured side. A 55.6 front shows amber on PSA 10. Re-check the line placement under the loupe, and price amber cards at the next grade down until a better photo says otherwise.
The PNG is one side of the corrected card, up to 1000 px wide, with the eight lines, L/R and T/B labels and a PRE-GRADE PREVIEW banner. The TXT has exactly three lines: L/R, T/B and the worst axis, each to one decimal.
Sixty percent is a starting estimate of how often a well-centered card with clean corners also comes back clean on surface and edges. Change it anywhere from 1 to 100 percent to match your own submission history; P10, Net EV and the decision update at once.
Yes. Upload the back and it gets its own set of eight lines, then both sides combine into one ceiling per company. PSA and CGC allow a 75/25 back for a 10, BGS 10 allows 55/45, BGS 9.5 allows 60/40, and SGC states no back figure.
Same engine