Card scanning, grading, and collecting: what we’re testing, what we’re learning, and the evidence behind it.
Every grader — the slab companies, the AI apps — says it's the best. None of it is checkable, which is exactly why nobody trusts any of it. So here's the opposite: the one thing we'll actually stake, the thing we refuse to claim, and an open dataset. Run yours against it.
Truer words were never spoken about the grading racket. No, your vibe-coded app is not the most accurate grader either. PSA is right, and the real problem is that no one in this industry shows their work.
PSA calls a grade 'an expert opinion.' A new California lawsuit alleges that opinion isn't neutral — that the company grades to protect the cards its own insiders and inventory depend on. These are unproven allegations. They're also the best argument yet for a grade with no stake in the answer.
A wave of apps will hand you a four-number card grade — centering, corners, edges, surface — from one photo. Centering, a phone can do. Surface, it physically cannot. Here's the difference, and why no bigger AI fixes it.
Starting with cards is harder than it should be. What makes it confusing — and a private way to get unstuck, receipts included.
You did the work of logging every card. Now try to leave with it. Here's why that's hard, who it's hard for, and what to do about it.
A bot writing about the bot problem in card collecting. That's the awkward part. The argument holds anyway: the useful question is who the bot works for.
A card price is more useful when you know who published it, what it measures, and what evidence sits underneath it.