Valdar / Free tools / PSA vs BGS vs SGC vs CGC vs TAG
Which grading company should you use? Weight the criteria yourself
Six criteria, five grading companies, and a ranking that shows its work: every per-criterion contribution is printed, so the result can be argued with rather than trusted. The trait values describe what each company concentrates on and how its holders behave in the market — they are not scores for grading quality.
Slide a criterion to zero to take it out of the ranking entirely. The weights are relative, so setting all six to 10 gives the same answer as setting all six to 5.
tied at 7.67 out of 10 on the weights you set
| # | Company | Sells easily | Label detail | Vintage | Modern | Slab look | Registry | Score | Gap |
|---|---|---|---|---|---|---|---|---|---|
| 1 | PSA | 1.67 | 0.33 | 1.33 | 1.50 | 1.17 | 1.67 | 7.67 | — |
| 2 | BGS | 1.17 | 1.67 | 0.83 | 1.50 | 1.50 | 1.00 | 7.67 | −0.00 |
| 3 | CGC Cards | 0.83 | 1.33 | 1.00 | 1.33 | 1.33 | 0.83 | 6.67 | −1.00 |
| 4 | SGC | 1.17 | 0.33 | 1.67 | 1.00 | 1.50 | 0.67 | 6.33 | −1.33 |
| 5 | TAG | 0.50 | 1.67 | 0.50 | 1.33 | 1.17 | 0.50 | 5.67 | −2.00 |
Each contribution is the company’s trait value multiplied by that criterion’s share of your total weight, and the six contributions add to the score. The trait values are descriptions of focus and market behaviour, not verdicts on grading quality, so this ranking is a function of the weights you chose.
| Company | Founded | Subgrades | Top grade | Standards | Fees |
|---|---|---|---|---|---|
| PSA Professional Sports Authenticator | 1991 | no | PSA 10 Gem Mint | grading scale | pricing |
| BGS Beckett Grading Services | 1999 | yes, four | BGS 10 Pristine, and BGS 10 Black Label when all four subgrades are 10 | grading scale | pricing |
| SGC Sportscard Guaranty Corporation | 1998 | no | SGC 10 Pristine | grading scale | pricing |
| CGC Cards Certified Guaranty Company (Cards) | 2020 | yes, four | CGC 10 Pristine, and CGC 10 Perfect when all subgrades are 10 | grading scale | pricing |
| TAG Technical Authentication & Grading | 2022 | yes, four | TAG 10 Pristine | grading scale | pricing |
Founding years and scales are each company’s own published description, read on 2026-09-11. Fees and turnaround times are deliberately absent: both change several times a year, so the links go to the live pages rather than to a copy that would be stale within a quarter.
What the table above leaves out
Those six columns are facts a company publishes about itself, and they settle fewer arguments than you would expect. A founding year tells you how long a label has been in circulation, which is a decent proxy for how many buyers recognise it and no proxy at all for how carefully the card in front of you will be assessed. A top-grade name tells you what the ceiling is called, not how often anyone reaches it.
What is missing from the table is everything people actually argue about: strictness, consistency, turnaround and price. Two of those change every few months, which is why this site links to the live pages instead of copying the numbers. The other two would need a controlled study — the same cards, submitted blind, to five companies, repeated — and no such study is published by anybody. Anyone ranking graders on accuracy is ranking them on anecdote.
So this tool ranks on what can be observed instead: where each company’s population concentrates, what its holder tells you, how the market treats its label, and whether a registry runs on it. Those are the six sliders above, and they are enough to produce a genuinely useful answer as long as nobody pretends they are measuring quality.
What the trait numbers are, and what they are not
Each company carries six numbers between 0 and 10 in this site’s reference data, and each one is an ordinal description of focus or market behaviour. SGC scores 10 on vintage because its population and its buyers concentrate there. BGS scores 10 on label detail because it prints four subgrades, which is a verifiable property of the holder. Nothing in the set is a verdict on how well any of them grades a card.
The consequence is that the ranking this page produces is a function of the weights you chose, and it changes the moment you move a slider. That is not a weakness in the tool; it is an accurate description of the decision. Someone selling modern rookies on a marketplace next week and someone assembling a vintage registry set over ten years are not asking the same question, and a single ranking that answered both would be answering neither.
Which is why every contribution is printed. Each cell in the result table is one company’s trait value multiplied by that criterion’s share of your total weight, and the six cells add to the score. If you think a trait value is wrong, the table shows exactly how much of the result it produced, and you can see whether correcting it would change the order at all. A recommendation between commercial services that cannot be checked like that is an advertisement.
Reading the six traits side by side
The shape of each company is more informative than its score. A profile that is high on two bars and low on four describes a specialist, and a specialist is the right answer when what you collect matches the specialism and the wrong answer otherwise. A flat profile describes a generalist, which is safer and rarely optimal.
Read the bars against what you actually do. If everything you submit was printed before 1980, the vintage bar is most of your decision and four of the other five bars are noise. If you buy modern parallels and sell them within the year, how easily the slab sells is most of it. The equal-weight default treats those two collectors identically, which is precisely why it produces a tie and why nobody should stop there.
Is there a best grading company?
No, and the equal-weight default on this page demonstrates it arithmetically rather than rhetorically. With all six criteria weighted the same, PSA and BGS tie exactly at 7.67 out of ten, because strength in one pair of criteria offsets strength in another. The tie is not a quirk of rounding; it is what happens when you refuse to say what you care about.
Move one slider and the tie breaks instantly. Weight how easily the slab sells and the answer goes one way; weight the detail on the label and it goes the other. The interesting output of this tool is not the name at the top but how far you had to push a slider to change it — a leader that survives every reasonable setting is a genuine recommendation, and one that flips at the smallest nudge means the two companies are equivalent for your purposes.
There are two situations where the question does have a clear answer, and neither comes from a ranking. If you are selling into a market where buyers ask for one label by name, that is the label, whatever the arithmetic says. And if a card is worth more in a holder it is not currently in, the question is not which company is best but whether the gap covers the fee — which is what the crossover calculator works out.
The tie at the top is the point of the page
With every criterion weighted equally, the top two companies score 7.67 out of ten and cannot be separated. One of them is strongest on how easily its holders sell and on registry support; the other is strongest on what the label tells you and on how the slab looks. Add those up with equal weights and they come to the same number.
That result is uncomfortable for a page that is supposed to recommend something, and it is the honest output of the arithmetic. Every ranking of grading companies you will read elsewhere has resolved this tie, and it resolved it by choosing weights on your behalf without showing you. Usually the unstated weight is resale liquidity, which is a perfectly reasonable priority and a terrible default for someone building a collection to keep.
So the instruction for this tool is the opposite of the usual one. Do not read the name at the top. Set the six sliders to describe what you are actually doing with the cards, then look at the gap column: if the leader is ahead by a tenth of a point, you have two equivalent options and should decide on fees or turnaround instead, and if it is ahead by two points, you have an answer worth acting on.
Which slider actually moves the answer?
That is a question with an arithmetic answer, and it is not the same for all six. Move one criterion from zero to ten with the other five left at the default, and the gap between the two companies that tie swings by 2.74 of a point for detail on the label and only 0.68 for how the slab looks.
Setting the weights for your own situation
- Start by zeroing what does not apply. A criterion at zero is removed from the ranking entirely, and most collectors can zero at least one immediately. If you have never entered a registry set, that slider is worth nothing to you and leaving it at the default is letting someone else’s priority into your answer.
- Decide whether you are selling or keeping. This single question moves the ranking more than any other. Selling within a year makes how easily the slab sells dominant; keeping for a decade makes it nearly irrelevant and promotes presentation and label detail.
- Weight the era you actually collect. Vintage and modern pull in different directions here, and splitting them is legitimate. Run the tool twice if your collection has two halves, and accept two different answers.
- Be honest about the label detail. Subgrades are genuinely useful on cards where one dimension is exceptional, and genuinely irrelevant on a card you are flipping next month. Weight it for how you use the information, not for how much you enjoy reading it.
- Read the gap, not just the order. The gap column is the whole result. Anything under a few tenths of a point on a ten-point scale is not a difference, and this page reports it as too close to separate rather than pretending otherwise.
- Then price the submission. The ranking chooses a company; it does not tell you whether to submit at all. That answer comes from the expected-value calculator, fed by the company’s own population report and its current fee schedule.
| Criterion | What it measures | Weight it high when |
|---|---|---|
| How easily the slab sells | How wide the pool of buyers is who will pay for that label without argument. | You intend to sell, and soon. |
| Detail on the label | Whether the holder tells you why the card graded where it did. | One dimension of your card is exceptional and you want it documented. |
| Suited to vintage | Concentration of pre-1980 cards in the population and of buyers who look there. | What you submit was printed before about 1980. |
| Suited to modern | Concentration of post-2000 cards, parallels and autographs. | You submit parallels, autographs and rookie-year cards. |
| How the slab looks | Matters more than collectors admit, and more for display pieces than for flips. | The cards go on a wall rather than into a box. |
| Set registry | Whether completing a registry set with these slabs is a thing people do. | You are building a set that a registry recognises. |
These six were chosen because each one can be checked against something a company publishes or against observable market behaviour. Strictness and consistency are not here, because they cannot.
| What you are doing with the cards | Sells easily | Label detail | Vintage | Modern | Slab look | Registry | Leader | Score | Gap |
|---|---|---|---|---|---|---|---|---|---|
| Selling modern cards within the year | 10 | 0 | 0 | 8 | 2 | 0 | PSA | 9.30 | +1.30 |
| Building a vintage registry set | 4 | 0 | 10 | 0 | 6 | 10 | PSA | 8.73 | +1.33 |
| Documenting one exceptional modern card | 3 | 10 | 0 | 6 | 8 | 0 | BGS | 9.15 | +1.26 |
| Vintage cards that go on a wall | 2 | 2 | 8 | 0 | 10 | 0 | SGC | 8.54 | +1.09 |
Set the sliders to any row and the panel above will print the same number. These are examples of priorities, not recommendations, which is why the weights are in the table: the point is that four ordinary collectors produce 3 different answers from the same six trait values, and that CGC Cards and TAG lead none of them.
What this tool does not measure
| Not in the ranking | Why it is not | Where to settle it instead |
|---|---|---|
| Grading accuracy and strictness | No controlled comparison is published by anybody. | Nowhere. A number here would be an opinion wearing a decimal point. |
| Fees and turnaround | Both move several times a year. | Each company's live page, linked above, then the cost calculator. |
| How your card will grade | That is a property of the card, not of the company. | Centering in the centering calculator; the rest is judged under magnification. |
| What your buyers ask for | It is a fact about one market, not about five companies. | The completed listings for the cards you actually sell. |
| Service, claims and disputes | Real and important, and impossible to score without inventing a number. | Absent rather than guessed at. |
Bottom line
Choosing a grading company is a weighting problem, not a quality ranking, and this tool treats it as one. The five companies carry six trait values each — how easily the label sells, what the holder tells you, vintage and modern concentration, presentation, and registry support — and every one of those is an ordinal description of focus or market behaviour rather than a score for grading quality. Set the six sliders to match what you actually do with your cards and the tool prints the full arithmetic: each company’s trait value multiplied by that criterion’s share of your weight, the six contributions, the total, and the gap to the leader. With every criterion weighted equally the top two tie exactly at 7.67 out of ten, which is the honest answer to “who is best” and the reason the sliders exist. Read the gap rather than the order, and if it is under a few tenths of a point, decide on fees and turnaround instead — both of which this page deliberately does not carry, because they change faster than any page can.
Questions this tool gets asked
Which grading company is best for sports cards?
There is no answer to that question in the abstract, and any page that gives you one has quietly chosen your priorities for you. What exists is a set of measurable differences: how widely a label is priced without argument, whether the holder explains the grade, which era a company concentrates on, and whether a registry runs on it. This tool asks you to weight those six things and then prints every contribution to the ranking so you can see exactly which of your own choices produced the answer.
Are the trait numbers quality scores?
No, and the distinction matters. They are ordinal descriptions of focus and market behaviour — SGC scoring 10 on vintage means its population and its buyers concentrate there, not that it grades better. None of the six numbers is a judgement about grading accuracy, consistency or strictness, because those would require a controlled study rather than an opinion, and no such study is published for any of these companies.
Why does PSA score highest on liquidity?
Because the pool of buyers who will price that label without discussion is the widest, which is a property of the market rather than of the grading. It is also the most cited reason collectors give for choosing a company, which is why it is one of the six criteria here rather than an assumption built into the result. If you are grading cards to keep rather than to sell, it is the criterion to weight at zero.
Does a BGS subgrade make a card worth more?
It makes a card explainable, which is not the same thing. Four subgrades tell a buyer why the card graded where it did, and on a card where one dimension is unusually strong that is worth real money — a 9.5 with a 10 for centering reads differently from one with a 10 for surface. Whether it lifts the price depends entirely on the card and the buyer, which is why this tool asks you how much you care rather than telling you.
Should I use the same company for vintage and modern?
Not necessarily, and this is the clearest case for weighting rather than picking. Vintage and modern have different buyers, different populations and different expectations about what a holder should look like on a shelf. Many collectors split their submissions on exactly that line and it is a perfectly coherent policy: run this tool twice with different weights, one for each half of what you collect.
What about grading fees and turnaround?
Neither is in this ranking, on purpose. Both change several times a year, and a number baked into a page is wrong within a quarter while still looking authoritative. Every company’s live fee page is linked in the table on this page, and the grading cost calculator turns a fee schedule into the all-in per-card figure that actually decides submissions, including postage in both directions.
What does a tied result mean?
It means the weights you set do not distinguish between two companies, which is information rather than a failure. With all six criteria weighted equally the top two on this page tie exactly, because one is strong on liquidity and registry while the other is strong on subgrades and presentation. Move any slider and the tie breaks immediately, which is the clearest demonstration available that the ranking is a function of your priorities.
Can I check the arithmetic?
That is what the contribution table is for. Each cell is the company’s trait value for that criterion multiplied by that criterion’s share of the total weight you set, and the six cells add to the score in the total column. If you disagree with a trait value, you can see exactly how much of the result it produced and by how much the ranking would change if it were different. A recommendation between commercial services should be arguable in that way.
The card is in your hand. Point the camera at it.
These calculators work on numbers you already have. Valdar gets you the numbers: it identifies the card from one photo, pulls what that exact card has actually sold for, and estimates the four grading sub-dimensions before you pay a submission fee. Free to try on iPhone and Android.