CGC Audit

[MatterEaterLad]

There's absolutely nothing to stop you, or a group of boardies, from doing this. I can't see why CGC would pay an external auditor when they already do an internal audit.

 

Your plan doesn't factor in damage to books in between gradings, by the way (either through shipping or some other way). I suppose your tolerable variance idea is supposed to cover that, but it might not in all situations.

 

Also, CGC would probably lose business if they graded the same every single time - the speculative resub market would disappear. The possibility of grade bumps is an important money-maker for them.

 

 

I've actually thought about doing it on my own.

 

After I submitted the Harlan books and the first 298 came back with 11 label errors (That's 1 in 27) I began wondering about human error in the process. And CGC fixed these right away. No complaint there.

 

I wanted to randomly resub 50 books just to track the variance before I ever do a big collection sub again. Just morbid, cynical, personal curiosity I guess.

 

I mentioned this to a couple boardies in confidence and they advised me that if I ever did this to show the results to CGC first -- that a poor report might disrupt a lot of things in the collecting community.

 

I always figured the variance would be in the single digits 2-3%. But Dan's JIM83 fiasco makes me think it's much higher, like 15%, which is why the resub game is so popular.

 

 

Don't whatever you do attempt to derail the gravy train.

 

Yep. That was the inference. That the profit margin between a 9.6 and a 9.8 is so fat that some people don't want any kind of change. They're happy to roll the dice on the resub game because sometimes it works in their financial favor.

 

 

 

Now sit down and be quiet.

 

 

"But...but...the Emperor has no clothes on..."

 

"Shaddap kid, I just resubbed a purple book and I bet big on blue!"

 

 

 

[vaultkeeper]

I would think that CGC would train their graders to grade according to their standards and probably test them before allowing them to grading on their own. What their allowable degree of accuracy is, I don't know.

 

But as the consumer, surely it's your opinion as to an 'allowable degree of accuracy' that's important?

 

So what is your opinion as to an 'allowable degree of accuracy'?

 

As far as the restoration check, I would expect them to get the correct color label every time.

 

 

I'm not sure how CGC goes about assigning a grade. Do the graders look over a book and give it a grade, or are they more precise than that?

 

If I were running a grading company and grading books on a 0-10 scale, I would have the graders grading on a 0-100 scale and down-grade a certain amount for each defect. Take off a certain amount for a 1/8 inch tear, more for larger tears. Same with other defects. The larger the defect, the larger the down-grade. If both graders are following the same criteria they should come close to the same number at the end of grading. If one grader ends up with a 72 and the other ends up with a 74 and the allowable degree of accuracy is 5, than both grades are within limits of each other and both are accurate results. I would average the two results and the final number would be 73 for a grade of 7.0 (no rounding up)

[Noodlez]

I audit the house before I yell at the wife.

[Sqeggs]

Here's a version of a point that someone made a while back that seems obvious once you think about it, but has stayed with me: Suppose by the randomness of the process a book that is actually (assume somehow we can tell) a 9.2 ends up with a 9.2 grade 50% of time, a 9.4 grade 25% of the time, and a 9.0 grade 25% of the time.

 

What would we expect to happen over time to high-dollar books? The answer, of course, is that the 9.0s and 9.2s get cracked and resubmitted until they come back 9.4s.

 

The conclusion seems to be that over time, the percentage of over graded books in the population of high-dollar books will continually increase.

 

 

[towards2112]

I audit the house before I yell at the wife.

 

If the business has 10 daily employees, and there is a problem, odds are

one of them funched up.

[MatterEaterLad]

For the record, I'm not anti-slabbing. I love graded books and have a ton.

 

And I can't imagine going back to a time where grading was so loose and restoration and doctoring (even by dealers) was so prevalent.

 

But before I'd ever do another massive submission, I'd resub 50 to get a sense at what I'm getting and whether I should pull the trigger or not.

 

If I ever do it, I'll let you know how it goes.

 

(thumbs u

[kav]

Perhaps when a book has an issue that there's no agreement on they just roll a 10-sided die and go with that...

[MatterEaterLad]

Perhaps when a book has an issue that there's no agreement on they just roll a 10-sided die and go with that...

 

You...you mean they don't?

 

 

[kav]

Perhaps when a book has an issue that there's no agreement on they just roll a 10-sided die and go with that...

 

You...you mean they don't?

 

THIS NEEDS TO CHANGE!

they need to start using a 20-sided die.

[MatterEaterLad]

Okay...not saying I'm going to do this, but if I were, what would be the best way to go about it without skewing the data? If I were to resub 100 books, should I...

 

Resub 100 random books?

 

Resub 100 books that are perceived to be over/under graded?

 

Resub 100 books and send them thru a dealer? Five different dealers?

 

Resub 100 books at once, or do 10 books each month?

 

Thoughts?

 

[Psithyrus]

Depends on what you're interested in checking. . .

 

If it simply a grade check, then all random.

 

If you want to do a resto check, as well, I'd consider that a separate "test" because it may skew the grading evaluation.

 

If you're checking that certain dealers get bumps, go through different dealers.

 

What exactly are you wanting to do? I'm not talking out of my rear, I've got experience in experimental set-up. I'll reiterate, that you may not need 100. Statistically significant results can be obtained from smaller sample sizes. We'd sometimes have 30-40 total. Saves money and time.