automated coin grading

I think the best we could ever hope for is automated analysis of details, nicks / bagginess, impact of minor problems (rim bumps), and luster. Maybe planchet porosity (EACs).

That's a pretty big head start.

I think it would ultimately lead to a continuum of grade, not the quantized system we have now. IOW, using all numbers. So you would have F-12, 13, 14, ... 19. Then VF20, 21, 22, ... 39. Then XF40...49.

Stick with integers, no need for decimals. The computer would compute a number like 48.783 and round to the nearest whole number.

The descriptors G, VG, F, etc would be despensible, but would stay around a lo-o-o-o-o-n-g time as a vestige.

 

Log in or Sign up to hide this ad.

I've not submitted a coin to a TPG and I have never owned one graded by a TPG either so you all will have to help me out. When a coin is graded by a TPG company the owner of the coin gets it back in a slab. On that slab is a grade. That grade is based on the condition of the coin, right? Where is the eye appeal rating on the slab?

The whole idea of the TPG was to be able to buy a AU-XX sight unseen and get a coin in a specific condition. This failed becuase it doesn't take into account eye appeal.

The person grading the coin may be biased for or against a coin based on how much that coin matches their eye appeal preference. Everyone has a conscious or subconscious "ideal" for what a coin should look like. That varies from person to person.

The visual grading guide doesn't seem to have any statements about "liking the looks of the coin so add or subtract points"

A coin in VF-XX can have better eye appeal than a coin in AU-XX, right? But one is still VF and the other is still AU.

Don't get me wrong about eye appeal being meaningless. Eye appeal has a huge impact on price and why I buy one coin over another but the grade is an absolute that is measurable and that assessment should be repeatable time and time again.

 

Andy

NOW , We are getting somewhere. I think what you are describing without trying to is the defferance between market and technical grading.

I believe a computer is fully capable of technical grading, providing a data base of charachteristics be human entered.

When we evaluate coins in slab, we are mentally market grading them ourselves without thinking about it.

Is this an average or a primo AU-55? Super strike? Mark free? Color? Is this worth AU-50 " bid " or " ask " or MS-60++ money ???

Case in point. I just sold a pair of AU-55 Lincolns yesterday, and on the 24-D, the buyer choose from 4 different coins, all graded 55 by either NGC or PCGS, then 2 graded 50 and 53, and all different. He chose, and paid more , for the best coin, my X-personal set NGC AU-55.

Eye appeal !!! All technically AU and all different to the human eye.

 

That's probably how the transition will occur someday. Maybe the coin will have to carry a split grade where the human disagrees or at least carry an explanation of why the machine was overruled.

 

The main difficulty is that there are no grading standards.
It's still all opinion.
And the opinions differ.

 

Well Jack, I only wanted to discuss automated grading and not eye appeal!!!!

Each company could have their own standard and let market forces determine the winner. In fact you could play the game that an AU-804.9987 from company A might upgrade to an AU-R.402 at company B. Nothing says the standards or the scale has to be the same between companies.

 

This one - "Photograde" by James F. Ruddy, 2005
page 84: Flying Eagle Cent About Good (AG-3) Obverse: The date will be weak but readable. Reverse. The rim will be worn down into the wreath.
page 89: Lincoln Head Cents 1909 to Date Extremely Fine (EF-40) Obverse: There will be wear only on the high points of Lincoln's head and face. Reverse. The lines of the wheat stalks will be very bold and clearly defined.

I don't see how any of the points you list as being a problem for some sort of computerized scanning process to pick up but I'm not an engineer.

Everyone is passionate about eye appeal but the point I've been forced to make is that in terms of grade (technical versus market a la Jack's response) eye appeal doesn't come into the picture.

 

Well I collect circulated coins so it is what I had lying around. Try this from
http://www.pcgs.com/grades.chtml


AU-50 Full detail with friction over most of the surface, slight flatness on high points
AU-53 Full detail with friction over 1/2 or more of surface, very slight flatness on high points
AU-55 Full detail with friction on less than 1/2 surface, mainly on high points
AU-58 Full detail with only slight friction on the high points

 

Yes
I believe a computer can technically grade coins

 

how would it be able to tell a weak strike MS from an XF-40 Lincoln cent???
I can see it now..
"grandpa, how did you grade coins when you were young?"
"well kiddies, we would hold them up so the light would hit them JUST right and grade it like that"
" but... wheres the computers? You had light back then?"

 

Andy,

I don't mean to be snide here but, to use one of my favorite quotes, "A little knowledge can be dangerous". What you seem to be doing is taking bits and pieces from grading guides and interpreting them literally. This is only going to lead to frustration and confusion for you.

What you need to do is to study grading. Take a course on coin grading. Read a book on how to grade coins. What I am getting at is you need to understand the concepts behind coin grading before you can begin to apply them.

You are trying to apply technical grading in a system that uses market grading. And in market grading eye appeal is HUGE. Eye appeal can take many forms - luster, strike, toning, etc. A coin that technically grades the same as another can be market graded higher than the other coin due to eye appeal.

Like others have said, some day computers may be able to consistently technically grade coins but I have serious doubts they can be programmed to consistently measure eye appeal.

I understand your confusion with market grading. I had to retrain myself to accept market grading. I am an engineer. To me 2 + 2 - 4, no if, ands or buts. It is absolute, not subjective. I applied the same logic to coin grading. I used to think a coin's grade should be based on what happened to it after leaving the press. In fact, that is exactly what I learned long ago.

My first grading guide - the 1970 edition of PhotoGrade - and the early ANA Grading Standards were based on technical grading. This was great for an engineer. But times have changed.

Somewhere along the line market grading entered the picture and that is how all the knowledgeable TPGs, dealers and collectors grade coins today. I had to retrain myself to be subjective when grading coins. Eye appeal is a huge component of that subjectivity. I am still on that learning curve.

You wouldn't attempt to fly a plane by buying a book, following the instructions on how to start the engine and operate the controls would you? (I hope not!) You would most likely take a course that teaches you a lot more than simply the mechanics of how to get the plane into the air. Similarly, a coin grading course will teach you more than simply measuring wear on a coin.

Do yourself a favor and educate yourself on coin grading. ANA offers an excellent home study course for a reasonable price. ANA also offers short courses in grading at their national shows. I'm sure there are other choices out there, too.

 

So there will be no bodybags in the future? If your coin is counterfeit it will simply be "destroyed"?!?! I don't know about you but I want my coin back either in a slab or in a bodybag, not destroyed.

And the grading services will be "factories"? Very interesting! Very interesting indeed!

 

With any luck, Andy Cook and " another prominent Lincoln guy " from here on the boards will be traveling to Long Beach with me. Nothing beats walking a large show, taking it all in.

Even if the budget is tight, the experience gained is priceless and contributes greatly to ones " arsenal " of coin buying/grading skills.

 

Hobo and everyone else, I appreciate your posts and contributions to the discussion.

My original question should have been about computers/technology doing technical grading. There are all sorts of technologies that could be applied. Maybe I should have said "assessing physical state" and not "grade" in the original post.

But the conversation turned down a different path and not necessarily a bad one at that.

I got my grading training and experience in the early 80's so I'm not hip to how things work now. When a coin is sent to a TPG and a grade is returned what is that grade? Technical or market? Or both? If market or both how does a TPG keep its grading consistent and repeatable?

 

One more problem with automated grading.
Not all coins in the same issue are graded by the same standards.
A 1877 IHC will never be subject to the same standards that most of the other IHC's would be.
Why? Because almost 100% of the 1877's were softly struck, so that eliminates using grading points such as the "diamonds".
If 1877's were subject to the same standards as most of the other IHC's, you'd have a hard time finding an MS coin.

 

Kanga, then the DB would have to have a larger reference set for grading differences like that. It could even extend to a reference set for each year, mint-mark, etc and so on. Storage is not expensive.

Thanks GDJMSP, this has resolved any doubts I may have had about TPG.

 

I agree.
In this case implementation would be a nothing more than a laborious job.

So the major fly-in-the-ointment goes back to my previous statement, there are NO universal standards.
And I see no resolution to that difficulty in the foreseeable future.