Written grading standards for uncirculated and proof coins can't/don't work...

[MarkFeld]

Many hobbyists talk about the need for, importance of and use of written/published grading standards. My view is that they simply can't/don't work for uncirculated and proof coins, though they might for circulated coins. I wish I could believe otherwise.

 

Please tell me why I'm right or wrong??

[MikeKing]

you're both right and wrong.

you're wrong because there has to be established criteriae

you're right because of the artistry behind determining a grade (i.e., evaluating all the nuance)

 

similarly, in medicine, you simply cannot publish the practice of medicine where you ultimately have to rely on your own gut feeling backed by tons of knowledge...that written knowledge being only a fragment of what it takes to determine a course of treatment, or in playing the violin, you can write a book on it, but there's no way on earth you could really describe how to play stacatto.

 

 

[James_OldeTowne]

Generally, I agree, except for the grades of MS-70 and PF-70.

[BillJones]

Generally, I agree, except for the grades of MS-70 and PF-70.

 

Funny I've seen some less than perfect PR-70 coins in my travels. So maybe they are not so easy to grade after all. (I NEED some gremlins here!)

 

But yes, I agree it is very hard to learn to get Proof and Mint State coins from words. If it were to be tried, perhaps color pictures would help. They would be far from perfect, but it would beat the words.

[MikeInFL]

Grading is subjective.

Subjective judgements are difficult if not impossible to verbalize with any accuracy.

Therefore you are correct, written grading standards can't be done.

 

That said, technical "sharpness" or "details" grades (and even perhaps "strike") can be documented in fairly accurate verbal detail -- it's the subjective nature of "luster" and/or "eye appeal" that are difficult to verbalize.

 

All IMHO...Mike

[IGWT]

I tend to agree with you. I especially enjoy the descriptions that make everything relative without giving a base point, e.g., a 64 has fewer hits than a 63, but more than allowed for a 65. Oh . . . ok.

[bsshog40]

I agree here also. The books I have are great for circulated coins but it has taken seeing alot of MS coins to be able to distinguish between those grades. When it comes to proofs, I really don't have a great idea of that process. I know when it can't reach 70 but 68 and below would be challenging as I have not seen much in those grades. MS and PF coins I don't think can be published as to grades for them. There are far too many reasons for one grade to another in these categories. Circulateds are easy, these mostly pertain to detail and the wear of details. MS and PF's are all based on contacts, blemishes, spots, slight abrasions, etc.... No way to publish a book that could properly and accurately explain these.

[dooly]

Hi all

 

check out Charles D. Daughtrey's Looking Through Lincoln Cents Chronology of a Series book

Pages 14-37 which is the "best" i have read Re grading

 

I used this book on my last lot of coins i sent in (cents) and i got them both spot on using his handy formular

 

sorry i cannot post more information as its copyrighted..

 

anyone intrested in grading should read this intresting book..

[Count_Erfit]

+1 for pictures. If the grading companies NEED grading sets, then take good pix and PUBLISH those.

 

 

[James_OldeTowne]

Mark, the funny thing is, I knew, I just knew you had the ANA grading guide in mind when you started this thread :crazy: !

[bsshog40]

Mark, the funny thing is, I knew, I just knew you had the ANA grading guide in mind when you started this thread :crazy: !

 

Photograde is another example james.

[rbrown4]

Many hobbyists talk about the need for, importance of and use of written/published grading standards. My view is that they simply can't/don't work for uncirculated and proof coins, though they might for circulated coins. I wish I could believe otherwise.

 

Please tell me why I'm right or wrong??

 

I guess it all depends on how one takes the phrase "don't/can't work". The replies above basically mirror my thoughts. Does the ANA standards provide one with a fool proof way to 100% accurately grade a coin? By itself, I would say no, is it a worthwhile guide or reference, I would say yes. The subjective part of grading will always be a case by case basis I believe. Written/published grading standards have served as a useful guide for me, but that is all that I expected of them, a guide.

 

Rey

[just-john]

Medicine music art take feelings knowledge understandings.But a coin is a coin it can be graded .You may have feelings for the coin history ect or it may have beautiful tones but the coin should reflect its grade.

[michael]

Generally, I agree, except for the grades of MS-70 and PF-70.

 

yes yes yes

 

 

[cpm9ball]

You're right to the extent that the guidelines cannot be etched in stone. It doesn't matter if you are referencing the ANA Grading Standards or Photograde or whatever. They all are just guides . As long as "grading standards" have a certain degree of subjectivity to them, they can't work.

 

If I were someone who was investing in coins for profit, it would scare me to death. That is why I'm just a collector. I have fun with coins and try not to concern myself with the opinions of others about what I should or should not collect. Beauty is in the eye of the beholder.

 

Chris

[DC Dawg]

Sometimes I feel I have an advantage in these types of discussions because of my technology and security background. I can see where stronger guidelines and standards can help make the process better for the collectors that will allow them to easily learn to judge for themselves without having to rely too much on "experts."

 

I have nothing against the experts and I have no arguments with the major grading services. But I think we all agree that there has to be a better way. Once again, this is where technology can assist.

 

Image analysis is used for everything from spying on the enemy to DNA analysis. Image analysis is used to look at chromosomes and can map out the genomic properties of these microscopic building blocks of life. From forensics to biometrics, image analysis is used to study the biological differences amongst us.

 

But coins are different. For the most part, coins are consistent. While VanAllen-Mallis has identified variations in Morgan and Peace Dollars, the Morgan Dollar is basically the same coin. Image analysis can tell the variety based on the baseline as can wear and luster.

 

Now I bet you want to know how this can translate to a non-electronic world. Other than producing high quality images with the narrative, this is one area where technology excels over paper. This is one area where technology can be used to benefit the industry.

 

Think about it... people carry around smart devices with cameras. "Smart phones" running PalmOS or Windows Mobile can be programmed to use the attached camera options to take a picture of a coin and run an analysis on what it sees. Specialized hand held equipment can do the same. Finer imaging systems providing a respectable one-time cost for the dealer and more serious collector can provide near instantaneous analysis.

 

I do not carry a smart phone for security reasons--they are a risk to some of the situations I have to work in. But if I knew a portable device was available for me to get an instant opinion about a coin that I could take to shows with me, I would buy the product in an instant.

 

I wonder if I can get VC money to do something like this?

 

Scott

[jesbroken]

I agree with you, Scott-right up to the part where the hand held equipment that has not been developed yet for coins, nor the software that has not been developed yet for coins. I doubt if this equipment and software could be implemented in the next 5 to 7 years. Unless, someone with a lot of money wanting to benefit the coin collecting industry and improve the lot of all collectors became involved. Maybe even a profit could be made from these devices. Coin theft identification would no longer be hit or miss. Oh well, just wishing, i guess.

[DC Dawg]

I agree with you, Scott-right up to the part where the hand held equipment that has not been developed yet for coins, nor the software that has not been developed yet for coins. I doubt if this equipment and software could be implemented in the next 5 to 7 years.

There is image recognition software that does work with deterministic information. I have worked on putting together image analysis systems (I do the systems work, not the image analysis). The software that exists now is amazing. From the information I have seen, I think a prototype could be produced within a year using one coin type.

 

I'm telling ya... if someone wants to provide some VC, I believe it can be done!

 

Scott

[MikeInFL]

Scott, There is a big difference between objective image analysis and subjective image analysis. So long as grading is subjective, computers will not play a major role in grading, and I challenge you to show an example where computers make subjective judgements with any accuracy. In short, I do not believe we are anywhere near as close to solving the grading problem with computers as you suggest, and the barrier is fundamental...Mike

 

[edited to add: That's not to say they couldn't help with the more objective part of grading -- for instance strike and wear -- but the subjective part of grading, eye appeal and to a lesser extent luster, is entirely out of reach with current computers and algorithms.)

[DC Dawg]

Scott, There is a big difference between objective image analysis and subjective image analysis. So long as grading is subjective, computers will not play a major role in grading, and I challenge you to show an example where computers make subjective judgements with any accuracy. In short, I do not believe we are anywhere near as close to solving the grading problem as you suggest, and the barrier is fundamental...Mike

A computer can measure luster and the number of marks on a coin. It can see doubling of letters when comparing it to a "perfect" image. If the strike is not strong, a computer can see the variation in the hairlines or clothes design.

 

Clear "S" versus filled "S" can be seen by a computer as a Micro O versus a regular O. Die clashes can be noted and VAM varieties can be determined from an analysis of the image.

 

Is that a wide "AM" on the reverse of the Lincoln Cent or a narrow "AM?" No matter how bad your eyes are, the computer can tell you.

 

Are those fields perfect or are there metal flow lines around the devices? The imaging system can tell.

 

What are you looking for in a coin that the right imaging system could not provide 95-99% of the information for you?

 

Scott

[MikeInFL]

What are you looking for in a coin that the right imaging system could not provide 95-99% of the information for you?

 

Only the most influential factor in determining the final grade (and to me most important) -- eye appeal. And therein lies the problem...Mike

[DC Dawg]

What are you looking for in a coin that the right imaging system could not provide 95-99% of the information for you?

 

Only the most influential (and to me most important) factor -- eye appeal. And therein lies the problem....Mike

I left you room for eye appeal. Remember, I never said computers would do 100-percent of the work, but I think they could do the vast majority of the work. Eye appeal is a human factor and not programmable--which does not bother me.

 

Scott

[MikeInFL]

What are you looking for in a coin that the right imaging system could not provide 95-99% of the information for you?

 

Only the most influential (and to me most important) factor -- eye appeal. And therein lies the problem....Mike

I left you room for eye appeal. Remember, I never said computers would do 100-percent of the work, but I think they could do the vast majority of the work. Eye appeal is a human factor and not programmable--which does not bother me.

 

Scott

 

First, I do not agree that eye-appeal is between 1 and 4% of grading.

 

Furthermore, I think that you will find that deciding the difference between grades from a wear and strike perspective is not quite as objective as you seem to think it is -- and this very issue part of what Mark was getting at in his original post. If we can't define grades objectively in verbal form, how are we going to teach computers the difference when we communicate with them verbally (i.e. through written programming languages).

 

That said, I think your 95-99% would be better estimated at 50%, and that is even a stretch, given the above...Mike

[DC Dawg]

That said, I think your 95-99% would be better estimated at 50%, and that is even a stretch, given the above...Mike

Like I said... reasonable people can reasonably disagree.

 

Scott

[MikeInFL]

That said, I think your 95-99% would be better estimated at 50%, and that is even a stretch, given the above...Mike

Like I said... reasonable people can reasonably disagree.

 

[Doing my best unreasonable person impression ] Since you just don't seem to understand the fundamental limitations of computers in making subjective judgements , I will try one last time to convince you with a different and much more practical argument...

 

Assume today that 4 graders and 1 finalizer each take 5 seconds to look at and grade a coin. That's 25 man-seconds total per coin.

 

Assume further with the introduction of the ScottB Grading Computer (SBGC), the sharpness, strike, and luster grade can be arrived at with 99% confidence (which I don't believe for a second, BTW).

 

However, the SBGC (by your own admission) is unable to take into account the subjective measurements of eye-appeal, and therefore humans must do this task.

 

Given the above, how much less than 25 man-seconds can the coin be graded for eye appeal and this factored into the SBGC sharpness/luster/strike grade?

 

Extending your argument above into man-seconds, you would say this goes down 95 to 99% -- less than one second. Using my argument, this number would drop in half, and come to think of it, I don't think it will save any time at all, as eye appeal has such an impact in grade (price) that the 4 graders and 1 finalizer will still be necessary.

 

In summary, taking the most rosy predictions that computers can grade sharpness, strike, and luster, the inability of computers to make accurate subjective judgements will yield a process that is no more efficient than that of today. In short, it will not save any time and therefore will have no effect on the bottom line of the grading company.

 

You can now have the last word.

[just-john]

So with 4 graders and 1 finalizer each take 5 seconds or 25 seconds per coin and there's

3660 secouds in 1hr that's 144 coins per HR at say $20 per coin thats only $2880.00

per hour or 1152 per 8hr shift 5760 for a 40hr week 23040 coins for a 160 hr month and they do what 100,000 per month so their's 21.70 people grading and finalizing.Now how good would your grading be after the first month computers don't get bored -blind -have off days

or days off.Computers grade fruit by color-people's face matches- check for voids in welds ect,Heck we put a man on the moon in 10 yrs.Don't think it would take 3 yrs if sombody wanted to do it

[carson city]

Why not ask the (CAC) COIN POLICE for guidance. Their arrogance is not to be believed in all things, coins !

[Zoins]

Can this argument be used to legitimize third-world slabbers?

[MarkFeld]

Can this argument be used to legitimize third-world slabbers?

Perhaps (but only) to a certain extent. However, despite the lack of widespread objective standards, at a certain point, the assigned grade on a coin becomes obviously too high in any knowledgeable person's view.

[Zoins]

Can this argument be used to legitimize third-world slabbers?

Perhaps (but only) to a certain extent. However, despite the lack of widespread objective standards, at a certain point, the assigned grade on a coin becomes obviously too high in any knowledgeable person's view.

But how does one define a "knowledgeable person"?

 

There's some precedence for the "reasonable person":

the reasonable person is appropriately informed, capable, aware of the law, and fair-minded. Such a person might do something extraordinary in certain circumstances, but whatever that person does or thinks is always reasonable.

Can we appropriate the reasonable person definition for coin grading with some modifications?