What are the statistics on proper grading by the two top TPG's?

[MikeKing]

I remember reading a book by..umm...that guy who's supposed to be a consumer advocate for coinee's? It's somewhere in my attic so forget about my looking for it, I went up there a couple of weeks ago for a copy of Breen and it was a disaster up there. Scott Travers! That's it! Scott Travers.

 

In this book, I remember him citing some 'study' that was done utilizing resubmissions and percent variance in the grading or something like that.

 

In my un-humble opinion, I think the top 2 TPGs accuracy in grading runs well below 90%, but >85%. And I am NOT kidding. *

 

Any opinions, ideas, data?

 

 

 

 

*disclaimer: this author is not necessarily supporting the notion that grade is the fools road to eye appeal

[WoodenJefferson]

One of the problem is, that we only hear about the bad returns from the grading companies, the should-have-been, but I was robbed, here, see for your self...only those who have sent in submission on a regular basis for say 5 years on a continuous basis would have a real grasp on the answer.

 

10% error in some industries would not be tolerable (say the airline industry) other commodities it's an acceptable percentage. 10% is feasible and so is 15% overall. Opinions are always subject to discrepancy, just like this opinion!

[michael]

show me a coin graded by pcgs and ngc and with michael photo seal for grade and eye appeal both services are 100%

 

unfortunately not many pcgs and ngc certified coins would pass my muster

[James_OldeTowne]

I think for market grading of generic series, their accuracy rate is 80%, and I'd go so far as to say that maybe 15% overgraded, 5% undergraded, the rest right on. For some series, the error rate is far higher, such as early copper (especially colonials). In fact, I probably disagree with fully NINETY PERCENT of the commercial grades on early copper.

 

For some series, such as Liberty nickels and three-cent nickels, I find PCGS, NGC and ANACS to be remarkably consistent and accurate. I'd peg them near 95%!!

[Oldtrader3]

Years ago, we did extensive testing on human inspection of medical devices. The inspectors probably spent about the same amount of time manually inspecting our products as TPG graders do grading coins. Inspectors were intensively trained for (8) weeks and had worked as assemblers for several years prior to becoming certified inspectors. To be certified, they had to repeat a controlled test which passed a final 80 sample per lot audit inspection by high level Quality Assurance Technicians (using 8X microscopes) for (3) sequential lots. Inspectors used 3X lighted magnifiers. Each of the lots they inspected contained 5000 units of production product with random defect levels.

 

The bottom line was that human inspection was <87% effective on finding significant defects even when inspectors were rotated every (2) hours. We reverified these inspection defect levels again when we installed computer vision inspection systems capable of indentifying <1 defect in 1,000,000 samples and the manual inspection closely correlated to our earlier 87% accuracy prediction.

 

Without getting into complicated statistical analysis, I would expect about <87% accuracy (within +/- grade), with some bias towards minus grades, when coin grading with experienced graders.

[EZ_E]

Usually, they're right on but there are definitely plenty of exceptions. Some toned coins may get a 2 point boost for the eye-appeal but, conserve that same coin and one may see suface damage from the toning. So, IMO, coins like this are overgraded and should be included in the statistics.

 

Key date coins are intentionally overgraded while common date are graded accurately. This is the biggest load of horse poopy that I can imagine. This places the buyer in double jeorpardy. Not only do they pay a huge premium for a not-so-scarce "key" date but they are paying much, much more for an overgraded key coin. (I mainly have the '16D Merc in mind)

 

Then there are coins that + or - depending upon what day of the week it is graded, if Jupiter is in alignment with Mars or if the grader got laid the night before. But this is all part of the wonder of man which makes us unique in creation.