Least reliable exactly where the money is

This is the most consequential misreading in modern collecting, and it is easy to make because the report looks exactly like an inventory. It is not one.

It counts submissions

Start with the mechanical fact. A population report records grading events, not objects.

If a card is graded, then cracked out of its holder and submitted again, it is counted twice. There is a procedure for avoiding this: send the old label back so the count can be decremented. Collectors who have tried report that the process is not well advertised and that they have stacks of labels they never returned.

The same card can also be counted by more than one company. A card that goes PSA, then to Beckett, then back to PSA sits in several databases at once, and none of those databases is talking to the others.

And the double-counting is not spread evenly

Here is the part that turns a technical quibble into a buying problem.

Nobody cracks a card out of its holder for fun. They do it when the gap between the grade they have and the grade they might get is worth the risk. Collectors who follow this closely describe the jump from 8 to 9 as the single biggest value creator in the crack-and-resubmit game.

So the resubmissions cluster at the top of the scale, on the expensive cards, at precisely the boundary where a buyer most needs an accurate number.

The rough consensus among people who have thought about it is that low grades are fairly reliable, commons are far more reliable than stars, sixes and sevens are better than eights, and vintage tens are reasonably accurate because there are few of them and they rarely get risked. The soft spot is the eight-to-nine band on cards worth enough to gamble with.

A card that reads as one of four PSA 9s might be one card submitted four times.

And for old cards it undercounts

The error runs the other way too, which is why "population report" is such a misleading name.

Enormous quantities of vintage material have never been graded at all. It sits in albums, in shoeboxes, in estates. A pre-war card with a population of thirty is not a card of which thirty exist; it is a card of which thirty have been submitted, out of an unknown number that survived.

We published figures recently that show the same effect from another angle. Among the cards this index tracks, the pre-1960 material averages PSA 3.75 with a single class grading 9 or better, while cards from 1980 onward average 6.53 with fifty-three. Some of that is genuine survival difference. Some of it is simply that nobody pays a grading fee on a card worth less than the fee, so the cheap end of every era is invisible.

What our own figures show about this

There is a pair on this index that makes the point cleanly, because it controls for almost everything.

Wayne Gretzky's rookie card exists as both an O-Pee-Chee and a Topps issue, same year, same number, same photograph. We track the same number of distinct grade and grader combinations for each of them: 116.

The average asking price on the O-Pee-Chee is more than three times the Topps.

The breadth of the graded market is identical and the value is not. Whatever the counts are measuring, it is not what the market thinks the card is worth, and it is not how many of each were made.

The figure that is actually useful

If you are going to use population data, the derived number is worth more than the raw one.

The gem rate is the count of tens divided by the total graded. Five hundred tens out of a thousand submissions is a card that grades well. Five hundred tens out of a hundred thousand submissions is a card that almost never does, and the second is the interesting one even though the top-line figure is identical.

Two other things change what the same number means.

How long the window has been open. A card issued in 2026 with three thousand graded is popular. A card issued in 1999 with three thousand graded, after twenty-seven years of people submitting it, is something else entirely.

And how fast it is growing. A population still climbing quickly is a population whose gem rate is falling while you read it.

What to do with it

Three things.

Read it as demand, not supply. A high total tells you a great many people thought this card was worth thirty dollars and a wait. That is genuinely useful information. It is information about the hobby's opinion, not about the printing press.

Distrust it hardest where it matters most. On an expensive card at the eight-to-nine boundary, the number is doing the least work and being asked to do the most.

And never treat a low population as scarcity on its own. Low population with no demand is not a scarce card, it is an unwanted one. The cards nobody submits are cheap for the same reason nobody submits them.

The thing worth sitting with

Somebody on a collector forum put it more bluntly than we would: the true population of a card will never be known.

That is right, and it is not a criticism of the grading companies. They publish an accurate count of the thing they can count, which is their own work. The error is entirely in the reading. We look at a number describing how often people paid for an opinion, and we see a census of what exists.

This index reports what is listed and what has verifiably sold, which are different measurements again with their own limits. You can read how we work the numbers out.

Sources