News that the Professional Coin Grading Service and Numismatic Guaranty Corp. have created a new “plus” grading system, and announced that they have computer technology to recognize resubmissions, is really the “Big One,” which David Hall, a co-founder of PCGS, promised would revolutionize the way that coins are bought and sold. It seems likely to.
Not only is the technology patented, or patent-pending, but what initially was supposed to be a PCGS-only announcement broadened to cover most of the industry with the addition of NGC. They officially added grades 63+, 64+, 65+ and more to the grading scale invented by Dr. William Sheldon as a pricing device more than 60 years ago when Harper & Row published Penny Whimsy.
Buried deep in the Web site was something more intriguing: the graders are not using a “+” or a minus symbol – that is computer generated – but are grading on a scale of 1 to 700, with uncirculated grades running from 600 to 700. This is a multiplication of 10 times the original scale of 1 to 70, with “1” being the worst-conditioned coin, and 70 the best. Proofs can be in any number, but are considered impaired if below -60, i.e., a Proof-45 1804 silver dollar.
Raising the bar to give a visual appearance of precision, the graders are internally learning to distinguish between, say, an uncirculated 640 (old MS-64) and uncirculated 650 (the old MS-65). If a coin comes out at 646 or higher, but less than 650, it gets the “plus” designation to the world, but the internal documents show that it is not 64+ but rather the three-digit integer that becomes its legacy number.
PCGS claims that its technology also allows for the capture of a coin’s “fingerprint” so that the addition of “putty,” or the use of jeweler’s rouge, or even some “handwork” designed to give the appearance of an improved grade, simply can’t be fooled once the coin has been scanned into the system.
That this is possible, I have no doubt, because it was something like 17 years ago that I saw Jim Diefenthal’s “Compugrade” in New Orleans with American Numismatic Association Executive Director Bob Leuver and comptroller Aimee Tihonovich. That technology allowed for Morgan silver dollars to be graded within two-tenths of a decimal point, i.e., MS-64.2, MS- 64.4, MS-64.6, and so on.
Inside a half day of training, I found that I could look at a coin without a magnifying glass and predict accurately to within 2/10ths of a grading point what the Morgan dollar would be graded by the machine before finalization. In full disclosure, ANA was looking to buy the technology but passed on it.
Continuing that this disclosure, what I was doing was not grading the coin so much as successfully predicting what the “grading machine” would view it as, and then writing it down to the nearest 2/10ths of a point. My accuracy was generally within a margin of error of 2/10ths of a grading point. I could tell, in other words, that it was a 64.4 and it might be a 64.2 or 64.6.
Having established that I was not a world-class grader, but could be taught to mimic the machine’s standards, the machine never was effective commercially because no market-maker was ever involved and did not figure out how to make the repetition commercially viable in a “break the slab” mold.
What is remarkable is that it became possible to grade uncirculated dollars to a degree of fineness previously thought impossible. Indeed, PCGS founder David Hall says human accuracy to a single point on a 700 point sale isn’t there, yet. It’s the use of computers that makes the process work.
Patents on the technology say that the 1991 Diefenthal “method and system for accurately and objectively evaluating the numismatic quality of a test coin and/or for fingerprinting the test coin for purposes of identification is disclosed.”
Specifically, the claim is made in the Diefenthal patent that, “Important to both the grading and fingerprinting aspects of the invention is the exact, numerical evaluation of any detracting marks on each side of the coin. In addition, systems and methods for illuminating an object surface with light at varying angles of incidence and for optically evaluating the object surface for features and defects, etc. are disclosed.”
The patent goes on to say that, “In a specific implementation of these systems and methods, the target object comprises a coin and the illumination and evaluation techniques are used to accurately objectively evaluate the numismatic quality of the coin and/or identify the coin. Important to the illumination and evaluation techniques is the ability to apply a uniform confined beam of light to the surface of the target object to be imaged.”
The inventors are listed as five Americans: Henry Merton, Jim Diefenthal, William Radian, Soumitra Sengupta and Emmet Lena Jr. The Application Number is EP19900110133, which was filed May 29, 1991. The patent’s Publication Date: 08/07/1991.”
There are 73 claims in the Diefenthal patent. Several of them worth listing before we even get to the 2010 PCGS methodology:
• An automated method for objectively assigning a numismatic grade to a test coin of particular issue
• Electronically identifying and locating each detracting mark on one of the obverse and reverse sides of the test coin;
• Electronically measuring the surface area of each identified detracting mark;
• Automatically assign to each identified detracting mark a quantity proportional to the detracting significance thereof based upon the location and measured surface area of the mark on one side of the test coin;
• To arrive at an amount representative of all identified detracting marks on said one side of the test coin;
• Automatically correlating said summed amount into a numismatic grade for said one side of the test coin with reference to a preexisting computer database of scaled values representative of numismatic grades;
• Electronically comparing the identified location and measured surface area of each detracting mark on each side of the test coin with a coin identifying computer database of detracting mark location and surface area information ... [against all] information in the database for a particular, previously recorded coin, thereby indicating identity of the test coin and said particular coin.
• Storing test coin detracting mark locations and surface area information in the coin identifying computer database for subsequent retrieval and comparison with other coins.
• Electronically microscopically evaluating the mint luster of each test coin
• According to “Streetinsider.com,” parent company Collector’s Universe “PCGS Secure Plus™ [is] a new high-tech grading process that has been designed to increase the confidence of collectors and dealers ...”
Besides the “new grade designation that potentially can increase the value of selected coins,” PCGS Secure Plus process uses laser scanning to help detect coins that have been artificially enhanced since their last certification and can also be used to help identify stolen coins.”
This means that the use of “putty,” the manipulation of metal to enhance grades may go the way of the do-do bird. A submission may still have its grade challenged – statistically, about three percent of coins graded by any grader of material submitted to any grading service result in an error of some kind – but not an artificial enhancement.
The error factor has to do with human nature and eyes. A great “world class” grader is right seven or eight times in 10. (By contrast, a collector of many years, like me might be “right” four to five times in the same 10. Given that three world class graders review the coin, a finalizer then melds them together.
In a world of 100 coins, grader #1 mistakenly under or overgrades 20 of them. Of these coins remaining, 80 percent are then correctly graded by grader #2, lowering the overall error percentage to about 4 percent. The third grader then judges 80 percent of that correctly (or about 3.2 coins) and the finalizer (who also makes mistakes) reduces it further to about 0.64 percent or about six coins per thousand.
That is in large measure a part of what drives crack outs – the hope that another look will yield a different result. Because sometimes it does, and the results can be startling.
From my vantage point, changing the grading scale by a factor of 10 is big news. (I thought the change would have been to adopt a 1 to 100 scale; but that evidently was rejected out of hand before the new internal scale was given a go-ahead.). It actually is designed to tone down what is essentially something controversial: the grade of a coin.
Grading has been controversial from the start. Consider this, from, The Numismatist, which is the monthly magazine of the American Numismatic Association (founded 1891). C.E. Bunnell, of Rochester, N.Y., in a letter, advocates a uniform grading standard:
“It is very important to members of the ANA that the Board of Governors take some stand with reference to issuing some kind of statement classifying coins so that all dealers that catalogue and sell coins ... must use the same classification. ... ”
He wrote that in a letter to the editor published in February 1913.
The Sheldon scale is just the latest in an attempt to quantify, and describe, the various states of preservation (or condition) that a coin is in. (Sheldon also tried to tie it to value with a -70 being 70 times the value of a coin in the theoretical basal state 1). It is not uniform; whether numerical descriptions or their adjectival counterparts are utilized, they differ from person to person, dealer to dealer, buyer to buyer, and seller to seller.
Most of the coin purchased by collectors, and the portion on which Sheldon spends so much time, are uncirculated. Yet widely respected commentators do not agree on how to describe differing types of uncirculated coins. Dr. Richard Doty, who wrote The MacMillan Encyclopedic Dictionary of Numismatics (1982), refers to “MS-65s, sometimes called ‘choice uncirculateds.’”
Scott Travers, the well-known writer and dealer (now a paid consultant to PCGS) terms “Choice Mint State-65” the “most frequently traded investor-quality coin,” utilizing the same standards that are employed by the American Numismatic Association’s grading standards, but then goes on to state that “Although ANA grading standards indicate that an MS-65 may be lightly fingermarked, the marketplace often dictates that MS-65 be virtually mark-free...”
In Mort Reed’s book, Coins: An Investors & Collector’s Guide (1973), James F. Ruddy was invited to contribute a chapter on coin grading and noted that, “Certain early coins are extremely difficult to find in Uncirculated condition without any bag marks or handling. Such flawless pieces are sometimes designated as Choice Uncirculated or Gem Uncirculated. Choice describes an above average Uncirculated specimen, well struck and with a minimum of minor bag marks or minting defects. Gem Uncirculated is the finest available, a sharply struck coin that is free of the usual minor bag marks or minting defects.”
In Penny Whimsy, (originally published in 1949, revised 1958) Dr. Sheldon said something that differed from all of this. “The MS-65 is a coin that would be a perfect MS-70 except for some small minor blemish. It may lack full mint luster, or some microscopic or almost negligible blemish may be demonstrable. There may be a spot of discoloration, a fingermark or a barely discernible nick.”
A later writer, James Halperin, in N.C.I. Grading Guide (1986) – probably the best technical book ever written on grading – notes that “choice” uncirculated is utilized by the ANA to denote coins that are MS-65 or proof-65, however he also notes that, “Many dealers apply the terms to MS/proof-63 and call MS/proof-65 coins ‘gem’”.
Thus, even the adjectives are not precise and have different meanings.
I always liked what my friend, veteran coin dealer Q. David Bowers, wrote in Adventures with Rare Coins (1979): “Often five different sellers will assign five different grades to the same coin, perhaps differing just slightly but still differing, often with important financial consequences ... As the evaluation of the grade or condition of a coin is a largely subjective matter, experts can legitimately differ ...”
The principal problem with numerical grading, of course, is that it gives the impression of exactitude and precision and it is, in the final analysis, little more than a subjective view of the cataloger. An article that I wrote, originally appearing in Numismatic News explained it thusly: “Standards, no matter how precisely defined, nonetheless remain subjective. Differences of opinion – not only between collectors and dealers, but even between dealers themselves, are bound to arise.” (“Under the Glass,” Sept. 20, 1975)
On Feb. 19, 1986, the ANA board, by resolution, approved the following statement, which I had a substantial hand in crafting: “Grading is an art and not an exact science. More precisely, grading is a matter of opinion. Differences of opinion may occur among graders as to a particular coin, and any grader could conceivably change his interpretation of the grading standards over the years.”
Not long after that, in a legal case in which I served as a consultant, the court observed the results of PCGS world-class graders efforts versus two witnesses hired by the Federal Trade Commission – two world class graders in their own right, Julian Leidman and Kenneth Goldman. They were given “test” coins to grade that already had PCGS designations.
The result: Leidman and Goldman graded the same 31 U.S. gold and silver coins and disagreed as to 19 of them, in five cases by more than two points.
That court also recited what happened when other experts also reviewed a series of coins that are generally conceded to be “easy” to grade – U.S. double eagles, a nice large, hefty coin where marks and scratches are relatively easy to see.
“One coin, a $20 gold piece, was assigned four different grades by five FTC experts, ranging from AU-58 to a ‘very nice’ MS-63. Another gold piece was graded MS-63 by one expert and MS-60 by another.”
There are those who claim that grading is synonymous with pricing, and that the marketplace pay identical prices for identically described coins. Well, it turns out that an arbitrary examination of this shows that that is not true, either.
Consider these examples of recent auction sales of certified coins involving a relatively easy to grade series, Morgan dollars.
1891-P Morgan dollar MS-64: Feb. 7, 2010 Lot 7963 PCGS MS-64, $747.50; and Lot 7964 PCGS MS-64, $546.25.
The price range: 36 percent on identically described coins, offered as consecutive lots.
Consider the same date, mintmark and grade (1891-P) Morgan dollar MS-64 as graded three times by PCGS and NGC once: Lots 8870 to 8874 had prices of $920, $575, $546.25 NGC and $517.50, respectively.
PCGS price range: 77.7 percent (not counting the different grading service).
Dr. Sheldon’s scale was initially based on numbers 1 to 70, which were intended by him to roughly correlate with some of the adjectival descriptions. However, unlike the adjectives, which referred to wear, Dr. Sheldon’s methodology was specifically designed to reference price. For example, a coin in Fine-12 condition was designed to be equivalent to one-fifth the price of an Uncirculated-60 coin, even though the wear on the coin is nowhere nearly as substantial as the numbers might suggest.
Under this scenario, MS-65 is valued at about 10 percent more than MS-60.
To briefly summarize the grading adjectival lexicon, coins may generally be termed poor, good, very good, fine, very fine, extremely fine, about uncirculated, or uncirculated, or defined as being struck as a proof or specimen coin intended for collectors.
Poor is the low end of the scale; uncirculated is the high end. Generally, the higher the grade, the greater the value for the same coin, and the more desirable to collectors. There are, of course, exceptions.
Although there were a number of adjectives that could be utilized to describe coins, the 1956 Guide Book of United States Coins (10th ed.) is typical of grading practices during that gestational period of time in the coin market: for most items, coins were priced in Good, Fine, and Uncirculated condition. Where examples were known, Proof was also listed.
To understand these three conditions of Good, Fine and Uncirculated, create a mental image of a 12-inch ruler, with the numbers running left to right. Generically, Good is at the left end, perhaps at the 1-inch mark, because it is a well-worn coin – of the type that had been in pocket change for a substantial period of time. The Fine coin is in the middle, perhaps at the 6-inch mark; its design elements are all quite clear, but it, too, has been in circulation.
A typical uncirculated coin is a “10,” with all that number connotes. The piece is not actually “un-circulated,” for if it had somehow never entered circulation, it would not have left the Mint. Rather, it never entered into general circulation for a protracted period of time. It still has its mint sheen, is bright, shiny, and with few contact marks (or bag marks) on its surface.
Imagine it to be similar to a brand-new copper “penny” fresh from the bank.
All other grades can be found along the continuum of the ruler. From 10 to 12 on the ruler are other better grades of uncirculated, which were verbally categorized differently decades ago than now, but still sufficiently different to understand that even 40 years ago, dealers and collectors distinguished between a typical uncirculated coin, and one that was better preserved or had fewer blemishes.
Today, thanks to the new David Hall announcement, that two inches between 10 and 12 inches has had a magnifying glass put upon it with numbers 600 to 700 found in-between. Using a yardstick, instead of a foot-long ruler, roughly every third of an inch is another grade: 601, 602, 603, and so forth. (From 64 to 66 [640-660] is just about seven inches on that yardstick.
How does all this look visually? Let me put it in a different context. The late Carroll Quigley, a historian at Georgetown University, once said that if you put every shade of human skin color next to each other, you could not differentiate and recognize one color from the other. Outer extremities would be very different, but not side by side.
That seems similar to what the case is with grading numbers. Its hard to see the difference between a 640 and a 641, but probably not as hard as a 640 with 645 or 646 (read 64+).
I asked PCGS if they could send me some photos of 1893-S silver dollars in MS-63, MS-63+, MS-64, MS-64+, and MS-65. Side by side, one grade from another, it’s difficult for someone who is not a world-class grader to tell the difference. But the “plus” is highly visible as a means of demarcation. See for yourself with the accompanying photos.
There’s a lot more to say about grading, how it has evolved and where it is going. For now, I’m content to sit back and let the marketplace speak its opinion – loudly– on what it thinks of the new grades of “+”, the internal grading numbers on the 1 to 700 scale, and machine recognition of images that prevents unwanted resubmits.