How Does PSA Measure Centering? Eye, Ruler or Technology

PSA does not publish what instrument its graders use to measure centering. What PSA does publish is the method: "PSA determines centering by comparing the measurements of the borders from left to right and top to bottom. The centering is designated as the percent of difference at the most off-center part of the card." That sentence, plus the per-grade tolerances and an explicit eye-appeal variance, is the entire official answer.

In practice, that means PSA centering is a human judgment applied to approximate thresholds — "approximately 55/45" for a Gem Mint 10 on the front, "approximately 60/40" for a 9 — with a published "5% leeway" for cards grading PSA 7 or better when eye appeal is good. It is not a machine-reported number, and you will never see a centering percentage on a PSA label or in a PSA report. That is the clearest difference between PSA and the two graders that do measure mechanically, TAG and ACE.

This article walks through exactly what PSA says it does, where discretion enters, why collectors see "inconsistent" results, how TAG and ACE differ, and how to measure at home in a way that mirrors PSA's published method as closely as possible.

What PSA officially says about measuring centering

Every sentence below is from PSA's grading standards page unless noted.

The method. "PSA determines centering by comparing the measurements of the borders from left to right and top to bottom." Two axes, each compared as a pair of opposite borders. Nothing about art, text boxes or the card back is defined here — the standards simply refer to "borders."

The worst-spot rule. Centering "is designated as the percent of difference at the most off-center part of the card." A card with a perfect left/right and a poor top/bottom is a poorly centered card for grading purposes. This is the same worst-axis logic TAG and ACE apply, just phrased less precisely. See our both-axes article for how this plays out on real cards.

The tolerances. PSA publishes a centering figure for every grade from 10 down to 1.5:

Notice the word "approximately" on every grade from 7 up, and its absence from 6 down. PSA's language gets softer exactly where the money is.

The 5% leeway. From the qualifier section: "A 5% leeway is given to the front centering minimum standards for cards which grade PSA 7 or better. For example, a card that meets all of the other requirements for PSA 9 and measures 60/40 off-center on the front automatically meets the PSA front centering standards for PSA 9. If a card meets all of the other requirements for PSA 9 and measures 65/35 off-center on the front, it may be deemed to meet the PSA front centering standards for PSA 9 if the eye appeal of the card is good."

The variance clause. Footnote on the standards: "At the grader's sole discretion, a small variance may be permitted on occasion based on the card's overall eye appeal."

The eye-appeal section. PSA states that "while most cards fall clearly within the centering guidelines for a particular grade, some cards fall either just within or just outside the printed centering standards" and that graders "reserve the right, based on the strength or weakness of the eye appeal, to make a judgment call." PSA's own example contrasts a 1955 Topps Koufax with a low-contrast yellow background, which may get an unqualified grade despite borderline centering, with a 1957 Topps Koufax with high contrast, where a card "that technically measures for a particular grade – let's say 70/30, may be prevented from reaching that unqualified grade because the market would view that card as off-center."

Put those pieces together and PSA's process is: a grader compares borders on both axes, identifies the worst, checks it against an approximate threshold, and then is explicitly allowed to move the result in either direction based on how the card looks. The full per-grade breakdown and the 60/40-to-55/45 history is in our PSA centering standards guide.

Human eye, ruler or technology: what we know and what we don't

The honest answer to "what centering tool does PSA use" is that no official PSA page names one. A few things can be said with confidence:

• PSA's standards are written in tolerances ("approximately 55/45"), not measurements. A grader working to that standard does not need a micrometer reading; they need to decide whether a card is inside or outside a band.
• PSA does not return a centering percentage with the grade. TAG does (on the DIG report) and ACE does (per-side percentages). If PSA graders record a precise number internally, it is not disclosed.
• The eye-appeal variance and the 5% leeway only make sense in a process where the grader, not an instrument, has the last word.

Collectors on grading forums sometimes claim PSA has moved to tech-assisted centering measurement. That is a collector claim, not something PSA has published, and it should be treated as unverified. It is equally possible that graders use a physical overlay, a ruler, a scan, or trained judgment — PSA simply has not said.

What this means for you: you cannot "out-measure" PSA. You can only measure accurately enough to know which side of a published line your card sits on, and then accept that within a few points of that line the result is a judgment call.

Why collectors see PSA centering results as inconsistent

Threads asking some version of "my card measured 54/46 and got a 9, a friend's 58/42 got a 10" are constant. Collectors report pixel-measured cards at 54/46 returning as 9s and cards near 59.5/40.5 returning as 8s; others report a scan-and-pixel-count measurement persuading PSA to bump a card on review. Treat all of that as anecdote, but the mechanics behind it are not mysterious:

Your measurement is not PSA's measurement. An angled phone photo, a sleeve shadow or the wrong reference edge can shift a reading by several points. On a card with roughly 2 mm borders, a 0.25 mm error moves the ratio about six points. A home reading of 54/46 can easily be a real-world 57/43. Our guide on how to measure card centering quantifies these error sources.

The worst spot governs. Collectors often quote the axis they measured. PSA applies the worst of left/right and top/bottom, front and back. A 54/46 front with a 78/22 back exceeds the 75/25 back tolerance for a 10.

The grade may not be about centering at all. PSA does not say which factor cost a point. A 9 with 54/46 centering may have a print line, a soft corner or surface wear you did not catch. The 2021 qualifier policy makes this more opaque, not less: PSA said the Off Center qualifier "will be assigned at the grader's discretion, and most typically will impact the numerical grade rather than carrying a qualifier." So a card that would once have come back "9 OC" now more typically comes back as a lower straight number with no explanation attached.

"Approximately" and "eye appeal" are real. PSA's own Koufax example says a card that "technically measures for a particular grade" can be held to a lower unqualified grade if the market would see it as off-center. The reverse is also in the text: a small variance "may be permitted on occasion." Two cards at 57/43, one with a thin low-contrast border and one with a thick high-contrast white border, can legitimately land differently under PSA's published rules.

None of that makes PSA random. It makes PSA tolerance-based with discretion at the margins. The practical consequence: a card that measures comfortably inside a threshold (say 52/48 and 53/47, with a 60/40 back) is a safe centering bet; a card at 56/44 is not a guaranteed 9 and not a guaranteed 10.

How TAG and ACE measure centering differently

Two graders have built their identity around taking the judgment out of centering.

TAG Grading says it uses "machine-learning image capture software which analyzes cards for defects," and that "all defects identified by TAG's grading system are thoroughly reviewed by human grading experts before Grade Finalization." Cards get a "TAG Score on a 1000-point scale" and a DIG report with "eight subgrades: front and back centering, front and back corners, front and back surface, and front and back edges." TAG's Pristine standard is stated as a measurement, not an impression: "Front Centering generally cannot exceed a single measurement of 51/49." Gem Mint is ~55/45 front with a back of ~70/30 (sports) or ~65/35 (TCG); Mint 9 is ~60/40 front. Details in TAG grading centering explained.

ACE Grading says centering "is calculated using 1/1000th of a millimeter accuracy measurements" via "digital measuring via microscopy," producing "a percentage for Left, Right, Top and Bottom centering for both the front and back of the card." Its 10 threshold is centering "less than a 60/40 split," and its OC modifier is rule-based (centering grade two or more grades below the overall grade). Details in ACE grading centering standards.

CGC Cards, for comparison, describes "multiple CGC Cards professionals examining every card aided by advanced technology" reaching "a consensus," but like PSA publishes no centering instrument or per-card number.

The trade-off is real. PSA's discretion can help a borderline card with good eye appeal and can hurt a card that measures fine but looks off. TAG and ACE will not rescue a 56/44 card with a nice look, and will not punish a 54/46 card with a harsh-looking border. If your card sits right on a line, that difference may decide where you send it. Our PSA 10 vs BGS 9.5 vs CGC 10 vs TAG 10 crosswalk runs that decision for common measurements.

How to measure centering the way PSA describes it

You cannot copy PSA's eye, but you can copy its published method exactly. The goal is a reading that tells you, with a margin of error you understand, which PSA band your card is in.

1. Take the card out of everything. No sleeve, no toploader, no slab. Edges under plastic read differently.
2. Get a flat, square image. A flatbed scan at 600 dpi or a straight-down phone photo in even light. Any tilt stretches one border and shrinks the opposite one.
3. Choose the reference edges PSA's wording implies. "Borders" means from the cut edge of the card to the outer printed border — the yellow edge on an English Pokémon card, the white edge on a vintage sports card. Do not measure to the artwork or a nameplate unless the card has no border.
4. Measure each border at two points (near each end of the side) and average. This catches diamond cuts and tilted prints.
5. Compute both axes. Left ÷ (left + right) and top ÷ (top + bottom), larger number first. Example: left 2.7 mm, right 2.3 mm → 2.7 ÷ 5.0 = 54% → 54/46. Top 2.1 mm, bottom 2.9 mm → 2.9 ÷ 5.0 = 58% → 58/42.
6. Repeat for the back. Blue Pokémon backs, text-heavy sports backs and TCG backs each have a symmetrical design element to measure to.
7. Take the most off-center reading — in the example, 58/42 front — and place it on PSA's table. 58/42 exceeds "approximately 55/45," so on published numbers this is a 9 on centering, with the 10 depending on the grader's variance clause.
8. Be honest about the margin. If your reading is within two or three points of a line, measure again with a better method, and assume the result could fall either way.

The free card centering calculator does steps 5 through 7 from a photo and maps the result to PSA, BGS, CGC, SGC and TAG thresholds; our grading standards guide has the thresholds side by side if you prefer to do it by hand.

Reading PSA's process like a grader

A few patterns follow directly from the published text and are worth internalizing before you submit:

• PSA 7 and above are where discretion lives. The 5% leeway applies only to PSA 7 or better, and "approximately" appears only on grades 7 through 10. Below that, the numbers are stated flat.
• The front is where the 10 is won or lost; the back is a gate. At 75/25 the back tolerance for a 10 is wide, but it is a cliff: 76/24 on the back is outside the standard regardless of how the front looks. See our guide on whether back centering matters.
• Half-points focus on centering. PSA says cards "that exhibit high-end qualities within each particular grade, between PSA 2 and PSA 9, may achieve a half-point increase" and that "there will be a clear focus on centering." A PSA 8 that measures like a 9 on centering is the classic 8.5 candidate.
• High contrast works against you near a line. PSA's Koufax example is explicit that a high-contrast border makes the same ratio look worse. A 57/43 thick white border is more exposed than a 57/43 thin pale border.
• Qualifiers are now mostly invisible. Since 2021, OC "most typically will impact the numerical grade rather than carrying a qualifier," so a centering problem usually shows up as a lower straight grade.

Bottom line

• PSA's official method is to compare border measurements left/right and top/bottom and designate centering at "the most off-center part of the card." It does not name an instrument and does not return a centering number.
• The published tolerances are approximate at the top (~55/45 for a 10, ~60/40 for a 9), with a 5% front leeway for PSA 7 or better and an explicit eye-appeal variance at the grader's discretion.
• "Inconsistent" results are mostly explained by home-measurement error, the worst-axis rule, non-centering flaws, and the discretion PSA writes into its own standard.
• TAG (machine-learning capture, human review, 1000-point score) and ACE (microscopy to 1/1000 mm, per-side percentages) publish measured, threshold-based centering; PSA publishes tolerances and judgment.
• To pre-check for PSA: card out of holder, flat image, cut edge to outer border, two points per side, both axes, front and back, worst reading against the table — and treat anything within a few points of a line as uncertain.

Frequently asked questions

Does PSA use a machine or software to measure centering?+

PSA does not publish what tool its graders use. Its grading standards describe the method, not the instrument: centering is determined by comparing the measurements of the borders left to right and top to bottom, designated at the most off-center part of the card, with stated tolerances per grade and an eye-appeal variance. Collectors sometimes claim PSA uses tech-assisted measurement, but there is no official statement confirming it.

What is the PSA 5% centering leeway?+

PSA's standards state that a 5% leeway is given to the front centering minimum standards for cards which grade PSA 7 or better. PSA's own example: a card that measures 60/40 on the front automatically meets the PSA 9 front standard, and a card measuring 65/35 may be deemed to meet it if the eye appeal of the card is good. It is discretionary, not automatic, at the outer edge.

Does PSA measure both axes or just the worst one?+

Both, but the worst one governs. PSA's wording is that centering is designated as the percent of difference at the most off-center part of the card. A card that is 52/48 left to right but 62/38 top to bottom is treated as a 62/38 card for the purpose of the centering standard.

Why did my card measure 54/46 but get a PSA 9?+

Three common reasons. The measurement may differ from PSA's because of perspective, sleeves or reference-edge choice, and a 54/46 home reading can be 56/44 in reality. The other axis or the back may be the problem. Or the 9 may have nothing to do with centering at all, since corners, edges and surface are judged at the same time and PSA does not publish which factor cost the point.

How is TAG or ACE centering measurement different from PSA?+

TAG states it uses machine-learning image capture software reviewed by human graders and scores cards on a 1000-point scale, with Pristine requiring a single measurement no worse than 51/49. ACE states it measures with digital microscopy to 1/1000 mm and reports percentages for every side. PSA describes a human comparison of borders against approximate tolerances with an eye-appeal variance, and does not publish a per-card centering number.

Can I measure centering the way PSA does at home?+

You can replicate the published method: measure opposite borders from the cut edge to the outer printed border, calculate left-right and top-bottom ratios on front and back, take the worst, and compare it with PSA's table. What you cannot replicate is the grader's eye-appeal call, so treat anything within a few points of a threshold as uncertain.

Sources & further reading

• PSA Grading Standards (centering, qualifiers, eye appeal)
• PSA Customer Update, June 2021 (qualifier policy)
• PSA glossary — Qualified card
• TAG Grading — Is TAG Grading AI?
• ACE Grading blog — Centering: An ACE Explanation
• CGC Cards — Grading Process