I’d pay an extra buck a card to change the frame color for sure. I bet people would pay 10% grading cost for higher end cards. Eg. 300 grade 30 slab color
am I ready for stricter grading? probably not. Maybe this will decrease the crying over how long it takes. I’m sure it will increase the number of people crying over low grades. “why did my card get a 2!?”
‘It will also provide unique card identification – or “card fingerprinting” – by identifying the exact card in order to track provenance, resubmissions, condition changes and other attributes over time.‘
Hello! (My first post. Let me know if this question should be asked anywhere else.)
Does anyone have experience with insurance on return shipping from CGC? I am about to submit an 84 card standard submission from the Netherlands and i am terrified about my Expedition set being lost in the mail with no financial compensation. CGC does not offer insurance on international return shipments but they say it is possible to use your own insurance policy. I am not sure how this works.
Is it normal to just accept the risk and use CGC’s standard policy for return shipping ($100 insurance per box… worthless) or is there something else i can do?
Also subgrades are increasing to $10 a card. The only memberships that will qualify for any grading fee discount are now at the Elite $299 tier. All current memberships will retain their original benefits until expiration/renewal.
This turns me off from PSA completely. Other than spotty service and huge turnaround times, who wants to grade with a company that will be as strict as possible with every little thing? Graders/buyers like us can’t see every single dot but a robot can. If a tiny surface mark that would have graded a gem mint 10/10 times in the past now becomes a 9 due to this change in the process then that would be ridiculous. I know I’m jumping the gun here obviously but it just seems crazy to me that grading, which has always been subjective is gonna be completely changed. I would’ve loved this if it was like this from the start… but at this point we’re at 6x certs almost. What does this even mean for the consistency/pricing of the cards that were graded before this change?
We dont know how the new grading system will grade yet. But I think its a good thing to try to take as much subjectivity out of grading as possible. Well see how its programed to distinguish something between a 8/9/10 but this very well could lead to newer certs obtaining a premium, with holders of older certs having an option to resubmit their old cards to see if they are in line with the new standards to obtain that new cert/case premium.
If the machine learning algorithm is trained from human-graded examples, it will also assign grades the same way a human would.
Modern ML algorithms are more comparable to how human brains work rather than how the strict and rigid classical computer algorithms work. Basically you need to “train” the algorithm with a huge number of real examples, for instance it could be a scan of the card and a human assigned grade as the correct “answer”. If you provide enough training examples, the algorithm is designed to recognize certain patterns and will effectively learn to assign grades based on those patterns. For instance, a properly trained algorithm will pick up on the fact that the whitening on the back edges of pokemon cards will dock the grade. The size, position and amount of whitening will be weighted by the algorithm to produce a final grade the same way a human would. The key difference is that you are not telling the program to LOOK for whitening, the program LEARNS that if it sees a certain threshold of whitening, a human would assign a lower grade.
The challenge of creating an algorithm like this is actually in curating the training data because the algorithm will pick up on any bias. For instance, if PSA grades charizords harsher than any other pokemon, a sufficiently complex algorithm will pick up on that and will also grade charizords harsher (for better or for worse). You also need a sufficiently broad and deep dataset to train on all the various styles of cards submitted to PSA or at least have enough data that the algorithm can learn very general concepts (ex. What “corner wear” looks like regardless of the cut, style or colour of the corner) that generalize to all cards
Just to get it out there as early as possible I want to place my next bets. On 3/12 I guessed:
CGC stands at the above quoted and PSA today is still $300+ lol. I honestly think PSA might not go sub $100/card this year, but wouldn’t be overly surprised to see $50-$75/card open by the fall. I predict CGC to not have an option under $30/card base and $40/card with subs by the fall. Giving a week notice (again) is a bold move because CGC is about to take in 6 months of work in the next 7 days probably. Will be interesting to see how this all affects the newest releases especially with a presumably popular Fall 25th Ann. release.
I’m a bit more interested in how the graded market will handle this. I assume the people who get their card graded to flip for money will just tack on the extra fee it cost to grade onto the card pricing. I wonder what that will do to the lower end of the graded market (cards worth 50-200 dollars). On the higher end I would assume it wouldn’t make that much of a difference.
Great news, I have to say I´m really pleased so far with what PSA did after they got acquired. Before, news and changes like these were pretty much unheard of - CGC also lit a fire under their butt I assume.
Looks like it could be a pre-grading tool users can use before submitting to get a better sense of expected grade. It’s probably going to take years of training data before an AI model is good enough to truly grade a card, I imagine they will start using this to detect centering/whitening and give cards a rough grade estimate that’s ultimately confirmed by a human grader.