PSA vs SGC vs BGS: does the grader change the price?

Sort any prospect card by sold price and the slabs arrive mixed together: a PSA 10, an SGC 10, a BGS 9.5, a raw copy. They look like four readings of one market. They are four different markets, and averaging them produces a number that describes none of them.

The grader is part of the card's identity

On this site a sale only counts as evidence for another card when player, product, parallel, print-run class, autograph status and the full grade string all agree — “PSA 10” against “PSA 10”, not “10” against “10”. A stated PSA 10 never matches a stated SGC 10, because those are different buyer pools with different price histories, and treating them as interchangeable is how a comp set quietly stops meaning anything.

That last row is a deliberate tolerance rather than an inconsistency. Some records were written before the grader was stored alongside the number, and refusing them would throw away real sales over missing metadata rather than over disagreeing metadata. Those are different claims, and only the second one is a reason to reject evidence.

The step between grades dwarfs the step between graders

Most of the argument online is about which grader carries the premium. It is the wrong first question. On prospect cards the gap between a 9 and a 10 from the same grader routinely runs two to five times — an order of magnitude more movement than any plausible PSA-versus-SGC difference on the same numeric grade.

So if you are checking a sold price and can only verify one thing about the slab, verify the number. A mis-read grade will cost you far more than a mis-read label.

The equivalence table we will not publish

The obvious move is a conversion: BGS 9.5 ≈ PSA 10, SGC 10 ≈ PSA 10 × something. There is no such table in this codebase, and its absence is a decision rather than an omission. Building one honestly needs a measured multiplier from a body of matched sales, and we do not have that body of sales. A ratio invented to fill the gap would be indistinguishable, on the page, from one we had earned — and it would silently price every graded card on the site.

There is exactly one place a cross-grader reading is allowed, and it is fenced off: when the real evidence ladder returns nothing at all, a fallback may look at the same numeric grade from a different grader at reduced weight, and the result is labelled an estimate that can never be called a bargain. Grade steps and raw-versus-graded are excluded from even that, for the same reason — no defensible adjustment exists yet.

What to do when you are buying

• Read the grader off the listing, not the number in the title. “10” on its own is not a comp; the slab label is the part that says which market you are in.
• Compare within one grader before comparing across them.If the only sales you can find sit in another grader's slabs, you have an indication, not a valuation — and it is worth saying so out loud before you bid on it.
• Never price a raw card off a graded sale. That is a bet on a grade you have not received yet, which is the break-even question rather than the value question.
• A population report is not a print run. How many exist and how many have been graded are different numbers, and only one of them is fixed.

Whether a raw card is worth submitting at all is a separate calculation, and it turns on the gem rate you would need rather than the one you hope for — the break-even calculator runs it. What a graded sale actually pays you after fees is a third question again.