How long before computerized grading?

[brothertona]

How many years until we have computerized grading, it would completely eliminate subjectivity.
A computer grading could judge centering perfectly, measure cards, could notice imperfections/alterations and could do this all in seconds.
What do ya'll think?

[ThoseBackPages]

8 to 10

[tjforce]

I highly doubt it would work.


Too many intricacies in cards. Can't see how you could standardize Chrome finish, foil finish, different colored boarders, etc.

A perfect example would be grading 1986 Fleer vs 2019 Optic or Chrome. The frayed-like edges of the 86 Fleer can still grade out a 10 since that's how they came out of the pack. Edges like that on Chrome or Optic would destroy the grade.


Plus if you could find a computer program that graded cards then it's most likely that someone can game that program and you're back to square one.

[hxcmilkshake]

I think someone out there with alot of money to blow could try to get something like that off the ground. Even if it couldn't do everything right off the bat but could still measure, catalog etc it would be good

Sent from my SM-N960U using Tapatalk

[ninjacookies]

Why would TPG's want to be 100 accurate and not subjective?

That would ruin their entire premise and gross profit margins. Got to keep them guessing and hooked like crack.

[ThoseBackPages]

Originally Posted by ninjacookies

Why would TPG's want to be 100 accurate and not subjective?

That would ruin their entire premise and gross profit margins. Got to keep them guessing and hooked like crack.

[ninjacookies]

Originally Posted by ThoseBackPages

[mvp]

We’re getting to the point where AI can do the work of radiologists because it can read scans better than humans. Grading cards is nothing compared to that. If there is money to be made, AI will be there.

[mondogenerator]

if it was there would be some bonuses, assuming they'd be part of the grading process - which they'd need to be:

recorded exact dimensions of the card at the time of grading - timestamped
image of front n back - so patch swapouts etc are easier to track
more robust database of cards so they can compare previously submitted cards vs original.

they may have to have any 10 personally sighted and maybe ditto anything under 7 to qual check.

the configuration for each variable would be tough - they'd need to almost build a template of each card design that's a perfect 10 and match to that.

it'd be crazy but what would be ideal is if Panini supplied a perfect example of each card to the graders as a benchmark.

im dreaming but, ya know.

else... on with the crap show.

[Josie27]

Asked this many times. Each card is designed to a specification. The artist draws it up a specific way. The boarders, the angels of each zig and each zag, specific colors for each team and so on. There are specifications fed into whatever program does the printing and cutting.

ARE THE SPECS FOR EACH CARD PRODUCED AVAILABLE TO GRADING COMPANIES?

How could grading be based on anything besides the degree to which it varies from manufacturer specification? This could include surface variations. Does the manufacturer believe the bubbles in Prizm are defects or part of the process? Each surface area had a color attributed. What percentage of the area is the color originally specified?

Cards are manufactured. Not handmade. You can measure defects in a manufactured product.

I get this will not work for cards from 1918 that were inside some cigarette package sent to American soldiers during the Meause-Argonne. Let's be real, unaltered cards of that age are the annommoly not the norm and make up a tiny % of the overall graded card market.



Sent from my SM-N950U using Tapatalk

[ninjacookies]

You guys are overthinking this.


TPG's are made specifically to have the human element of error, bias, and subjectivity.

Their profit model is based on having people submit the same card 3 times with up to 3 different grading outcomes. The resub and BGS .5 from gem game is real.


'I'm stumped. It looks visually perfect, I swear I combed over it 5 times with a jewelers loupe...psa robbed me on this one...definitely sending it back in.'

[TimBuckTwo]

It would sure pick up trim jobs

[ThoseBackPages]

would the robot damage the card?

[RogerGodahell]

I would be happy with a photo registry for graded cards.

[cking]

Originally Posted by ninjacookies

'I'm stumped. It looks visually perfect, I swear I combed over it 5 times with a jewelers loupe...psa robbed me on this one...definitely sending it back in.'

The funniest part about it is that these people sending cards in are not smart enough to realize that PSA does this on purpose to them.

I mean they really dont get it. Its amazing. They are the quintessential rubes.

[BlowoutBuzz]

I wouldn't assume the current companies to be on the forefront of technology. It would likely be a new entity.

[oldgoldy97]

Originally Posted by mvp

We’re getting to the point where AI can do the work of radiologists because it can read scans better than humans. Grading cards is nothing compared to that. If there is money to be made, AI will be there.

[ballhawkdawk]

Originally Posted by ninjacookies

You guys are overthinking this.


TPG's are made specifically to have the human element of error, bias, and subjectivity.

Their profit model is based on having people submit the same card 3 times with up to 3 different grading outcomes. The resub and BGS .5 from gem game is real.


'I'm stumped. It looks visually perfect, I swear I combed over it 5 times with a jewelers loupe...psa robbed me on this one...definitely sending it back in.'

All of this. I personally can't fathom sending my cards to some dipsh1t to tell me whether or not it's in good condition. I have eyes.

[patchgenie]

robby the robot might be better than the people they have grading now.

[mondogenerator]

i dont know if we need full AI grading, but a good first step would be to include basic tech like measurements, photos, timestamps, recording serials for comparison etc

[whoshotjr]

I think there could be two types of services.

#1 - Computerized version of the current system where card gets put in a holder

#2 - Some sort of app/subscription for analyzing cards using a phone.

The second probably wouldn't handle surface issues, but it could give a quick/dirty analysis of card size, centering, corner sharpness, etc. Maybe tie it to a photo-sharing service or provide some sort of digital score on the photos. More of a show-off my collection service versus something for authenticating/selling.

[3rdpedal]

i work in manufacturing (not printing) and more specifically the IIOT (industry 4.0) field of AI/ML/Sensor enablement. From my experience, this would be costly to develop yet incredibly accurate and well worth the investment from TPG. I hear what some of you are saying about the TPGs not wanting to be accurate, but assuming they do for a second, this is the perfect application for a vision system.

[Patsguy5000]

Interesting, seems like it might catch altering/trimming and maybe more consistent.

[hermanotarjeta]

Never - no company would be dumb enough to do this.

How many return customers would you get if you were giving out 7,8 and 9’s, while your competition is giving out 9.5 and 10’s?

Today’s grading industry cares nothing about the actual cards themselves, they only care about the business of grading cards.

I’m still waiting for the first company to give out the first “11” grade - that’s where the money is at.

[Deadshot]

Originally Posted by hxcmilkshake

I think someone out there with alot of money to blow could try to get something like that off the ground. Even if it couldn't do everything right off the bat but could still measure, catalog etc it would be good

Sent from my SM-N960U using Tapatalk

I've heard of a couple of people working on something like this, but not any of the major grading companies. So I think the initiative is out there. Who knows if we'll ever see it though.