Non-subjective grading system

cjmbuffalo

Would collectors except a electro-mechanical grading system that is devoid of human subjectivity? If the standards of coin grading were to be reduced into areas that could be measured in quantitative terms, thus eliminating any subjectivity and "current pricing" thinking would this be better and more consistant then what we have today?

LucyBop

To simplify the above statement in laymens terms, 'Hey HepDaddyHall, just grade every Frankie Lucy sends a ms66fbl and that'll be outa sight Hepcat!'

shylock

PCGS experimented with this over a decade ago. It seems like this sort of grading could work for the 80% of a coin which can be technically analyzed, and then let the consumers decide with their pocketbooks on the other 20% that's more subjective and falls into the eye appeal category. But then coins would never have a chance of upgrading unless the software was altered, and that would be a financial nightmare for PCGS. They may not have considered that drawback back in 1990 but my guess is once they did, it abruptly put an end to this experiment.

cladking

If absolute standards were defined then humans could grade coins with
some consistency. One would have a good idea of what a coin looked
like by it's grade.

GDJMSP

I can only hope, wish & pray for the day when all grading companies follow the same & consistent set of standards when it comes to grading coins. I cannot see a computer doing this because not every coin can be graded in the same manner. There are just too many variables for different denominations, dates and mints - but I can see a human doing it.

Of course if they did and the coin was graded accurately to begin with - upgrades would no longer be possible.

leothelyon

GDJMSP
Welcome to the boards.

Of course if they did and the coin was graded accurately to begin with - upgrades would no longer be possible.

Because the technology already exists, I believe such a system is possible. And upgrades would be possible because the information that is fed into the computer, can also be changed. Sure the mixing of color, luster and strike may proved to be impossible for a computer to analyze but that's what man can grade, the eye appeal. Someone does have to feed the machine anyway. Just think of it, a low grade MS63 coin with high end eye appeal. Where do we go from there? lol

Leo

dorkkarl

it ain't possible w/ today's technology - at least not to the precision that would exclude all room for debate.

K S

nwcs

Never happen so long as:

1. eye appeal remains a factor.
2. weak strikes are not properly accounted for
3. the current system is adequate for the market's purposes
4. a machine cannot detect all the faults of a coin

cjmbuffalo

I do believe that it is possible to create a system. As long as the system does not use subjective standards. Everything else can be measured, thru some type of computer program. Laser scanning today can resolve incredibly small detail and then convert to a 3D photograph. The subjective side of the equation (such as color) could still be used but maybe with less contribtuion to the overall grade.

Getting the individual out of the process is the only way to create consistency.

cladking

A descriptive grading system could easily be instituted. The technology does not yet exist
to use this system, but a person could use it. Eye appeal is for the main part just a sum-
mation of the parts of the grade. If collectors demanded true standards for grading such
a system would evolve. Instead we accept the current system because we think it makes
pricing easier. In actuality we know neither the grade nor the value for most coins.

dorkkarl

<< <i>I do believe that it is possible to create a system...Everything else can be measured...Laser scanning today can resolve incredibly small detail and then convert to a 3D photograph. >>

that's not the problem. the problem is that the algorithm to program such a perfect grading system would be enormously process-intensive, ie. for a single mark, consider the factors involved: width of the mark, it's depth, shape, disturbance of the surface, reflectivity, chemical reactivity, etc, then multiply the # of factors TIMES the proximity of every other mark & tone-spot on the coin's surface, and the computing power necessary becomes phenomenally complex.

that's just the tech. aspects. the REAL problem is that grading is supposed to establish the relative value of coins. but grading is NOT the only, & certainly not the most important aspect of a coin's value. the desire of an individual to purchase the coin is. so even if such incredible precision were possible technically, it would not amount to much advantage, since (although most folks will probably not want to believe this) grading IS NOT the most important factor in a coin's value.



<< <i>The subjective side of the equation (such as color) could still be used but maybe with less contribtuion to the overall grade...Getting the individual out of the process is the only way to create consistency. >>

exactly



<< <i>A descriptive grading system could easily be instituted >>

we already have 1!



<< <i>Eye appeal is for the main part just a summation of the parts of the grade. If collectors demanded true standards for grading such a system would evolve. Instead we accept the current system because we think it makes pricing easier. In actuality we know neither the grade nor the value for most coins. >>

if the object was to sell coins to a computer, the purely objective solution might work, but the point is that your trying to remove a factor from the "grading scenario" that integral to the scenario itself. the fact is that humans are buying the coins, so by definition, it is NOT possible to remove the human factor. nor should you want to for that matter.

think of it as being like trying to remove the "human " factor from a football game. if you did that, footbal would no longer be a human endeavor. you would just play the season out on your laptop at home.

K S

Cam40

I think we do have the technology to create a program that takes into consideration nearly ever aspect of a coins features.In fact,I beleive we have the technology to counterfiet coins to such detail that they are virtually indistinguishable from authentic pieces.Once this is `discovered` all hells gonna break lose in the coin collecting world....
Someone had to say it...AGAIN..

perfectstrike

I can't resist. The "technology" does exist. That's the easy part. The algorithms are being developed and refined. The mathematics is being used in more and more programs for various hard sciences. Computation power isn't really a factor as it was even 5 years ago. The limitation at the moment appears to be accurately graded coins to be fed into the system to yield comparison tables. Given that none of the grading systems, and companies sing them, have extensively published criteria regarding mark locations, size, depth, shape and other factors, a large sampling is required for a system to become "smart" enough to distinguish between grades.

Many factors such as lighting, color and varieties can be accounted for to some degree, if all the rest were knowns. There are a couple of projects out there producing interesting, if limited, results.

perfectstrike

cjmbuffalo

Information regarding the technical parts of the coin, size, depth, etc can be obtained from the Mint under the Freedom of Information Act. This would give the benchmark for the system to measure against.
This also would just about assure the ability to detect a fake or altered coin. The only area that cannot be measured is the "subjective" like eye appeal, color and how that plays to the grading of the coin. It would seem that the grading could be done on a technical basis. Establishing a technical grade would just be a part of the value pricing, as the owner then could work on the subjective part of the pricing based on eye appeal and supply.

MrKelso

WIth laser technology it could be done, But the cost of developement and the final manufacturing cost of the machine would be a GAZILLION Dollars. We get stuck with the bill.