Measuring Eye Appeal of MS Toned Morgans
gemtone65
Posts: 901 ✭✭✭
In a recent thread by Anaconda on grading levels, I suggested that the absence of a measure of eye appeal is a more important issue than refining the grading scale. Perhaps in response, Adrian started another thread on eye appeal. Unfortunately, that thread quickly deteriorated into a discussion of pretty girls and tawdry looking men. So, I offer this new thread in a serious manner.
To make this as brief as possible, let me offer a proposal for developing such a measure, and ask for comments. Suppose I (or anyone so inclined) took a large sample of toned Morgan dollars and simply ranked them from "best'" to "worst", not on technical grade but on "eye appeal." Thus, if I were unfortunate enough to own one of the toned Morgans offered on E-bay by, say, Blue Moon Coins, such drab brownish-black pieces would be near the bottom of the ranking. And, if I were lucky enough to some coins formerly the property of, say, GSAGUY, Art R., Anaconda, or the Goldberg's "toned dollar collection" (which in fact I am), then these pieces would be up among the very highest rated coins.
Now, after making this assessment, and assigning a rank to each piece, suppose I then identified the charateristics that separated the best, from the middle of the pack pieces; and, those in the middle from the worst. thinking about it, such charateristics include the following: number of different colors; types of colors; location of colors; and, proportion of coin that is toned. With a little effort, these charateristics can be quantified. And, they are objective, i.e., we could all agree on how to rate those charateristics once the rating scale is specified.
Notice, I left out luster as a charateristic. I did so not necessarily because I thik its approporiate to do so, but because 1) its effect on the coin may already be included somewhat in the grade, and 2) it is difficult (but not impossible) to measure luster objectively. So, whether it should be included, and ifd so how it can be measured, is something we should discuss, along with other important charateristics that I have omitted and can be measured objectively for the most part.
Once the desired charateristics have been identified, and a numerical identifier created for each such charateristic, I would evaluate each coin in the sample with the intention of coding each characteristic. Then, I would use the data and undertake the needed statistical tests that would generate a weighted index, in the form of an equation, that would allow me to predict the rating of a coin given an objective measure of each charateristic. How well the equation actually predicts eye appeal would depend, in part, on the correlation beyween the predictions of the derived equation and the actual ranking itself. There are statistical tests that provide the basis for making such assessments.
In effect, what this type of experiment would do is provide a weight for each charateristic as it affects eye appeal. And, if there was general agreement about what is attractive and what is not, and on the most important factors affecting eye appeal, then this equation could be used by almost anyone, such as PCGS or NGC, as a measuring index for eye appeal.
Comments?
To make this as brief as possible, let me offer a proposal for developing such a measure, and ask for comments. Suppose I (or anyone so inclined) took a large sample of toned Morgan dollars and simply ranked them from "best'" to "worst", not on technical grade but on "eye appeal." Thus, if I were unfortunate enough to own one of the toned Morgans offered on E-bay by, say, Blue Moon Coins, such drab brownish-black pieces would be near the bottom of the ranking. And, if I were lucky enough to some coins formerly the property of, say, GSAGUY, Art R., Anaconda, or the Goldberg's "toned dollar collection" (which in fact I am), then these pieces would be up among the very highest rated coins.
Now, after making this assessment, and assigning a rank to each piece, suppose I then identified the charateristics that separated the best, from the middle of the pack pieces; and, those in the middle from the worst. thinking about it, such charateristics include the following: number of different colors; types of colors; location of colors; and, proportion of coin that is toned. With a little effort, these charateristics can be quantified. And, they are objective, i.e., we could all agree on how to rate those charateristics once the rating scale is specified.
Notice, I left out luster as a charateristic. I did so not necessarily because I thik its approporiate to do so, but because 1) its effect on the coin may already be included somewhat in the grade, and 2) it is difficult (but not impossible) to measure luster objectively. So, whether it should be included, and ifd so how it can be measured, is something we should discuss, along with other important charateristics that I have omitted and can be measured objectively for the most part.
Once the desired charateristics have been identified, and a numerical identifier created for each such charateristic, I would evaluate each coin in the sample with the intention of coding each characteristic. Then, I would use the data and undertake the needed statistical tests that would generate a weighted index, in the form of an equation, that would allow me to predict the rating of a coin given an objective measure of each charateristic. How well the equation actually predicts eye appeal would depend, in part, on the correlation beyween the predictions of the derived equation and the actual ranking itself. There are statistical tests that provide the basis for making such assessments.
In effect, what this type of experiment would do is provide a weight for each charateristic as it affects eye appeal. And, if there was general agreement about what is attractive and what is not, and on the most important factors affecting eye appeal, then this equation could be used by almost anyone, such as PCGS or NGC, as a measuring index for eye appeal.
Comments?
0
Comments
I think that this is a great test case for an individual grader who has mastered a high level of consistency in determination of eye appeal within a series. In other words, I would bet that you could come up with the algorithm needed for your own collection and could rank the coins and then look at them next year and rank them again to come with much the same numbers. Although most people will generally agree that certain colors or patterns or coin coverage are more desirable than others you will still encounter a fairly high degree of collector bias for certain characteristics that might not be able to be overcome. Also, if my memory serves me correctly here, I believe Iwog stated about a year ago that he was aware of a scientific system that could accurately and reproducibly measure luster and that he had been using it for quite some time. I asked for the details but never received an answer. Your proposed format for the equation reminds me of advanced weather prediction formulas, which normally breakdown over a short period of time, however, we are not talking about a dynamic system with numismatics. We are really looking for what are the weighted values for each characteristic, what are the pertinent characteristics and what are the rankings for various "looks" within each characteristic. We might call it the "gemtone65 Eigen value". I kind of like the way that sounds.
In honor of the memory of Cpl. Michael E. Thompson
At any rate, I'm really glad you like my ideas, and perhaps almost as importantly, that you cared to respond. One aspect that I kind of like about my suggestion is that it doesn't require anyone to value individual attributes; all they have to do is distinquish those attributes or characteristics in some systematic way, and then express that difference quantitatively. Then, fitting these quantified attributes to the original ranking REVEALS the relative importance of each attribute, and simultaneously allows identification of the relative eye appeal across a broad variety of toned Morgan dollars.
In the end, you could wind up with the following results: say my sample actually involves only my own gem toned morgans, with values (i.e., market prices) varying from twice bid to 16 times bid. Suppose I derived an eye appeal index that arbitrarily is set to vary, say, from 0 to 100, and used my coins for construction. Now say someone plugged into my eye appeal formula the charateristics of his toned dollar, which perhaps involve 50% toning, is monotone, has light blue color, that color is located from 11:00 to 5:00 o'clock, and luster is "above average."
And, let's say he finds that the index computed is 90. This could be interpreted to mean that his coin would be in the 90th percentile of coins of comparable toned quality to mine, i.e., better than 90% of my coins. And, any third party who had a sense of what my inventory looked like would understand what it meant to have such a coin -- even GSAguy would like it.
The problem is that describing it to allow someone else to relate it quantitatively is pert near impossible.
The coin had intense colors.
How intense?
Well really intense, as intense as coin #6 on the Gemtone65 scale.
Oh, I didn't think coin #6 was that intense.
What kind of monitor are you using?
A Shebornic 6600.
Really? What is the contrast set to?
Blah, blah, blah.
So, did i understand your original post correctly?
I like four levels of eye appeal.
1) Ho Hum.
2) Nice.
3) Sweet!
4) I'd leave my wife for that coin.
adrian
adrian
Also, in my last response I forgot to finish my point about the relationship between rankings and price. Recall I said that the pieces in my collection have values that range from twice to 16 times bid. Well, once you have the ranking, or even better the eye appeal index measure, you can then correlate it to the value in terms of multiples of bid. This would allow someone to compute a fair price for a given coin.
For example in the case where the index showed the coin to be a "90", you could find the expected multiples of bid associated with such a coin. Based on my sample, a "90" might convert to a value equal to 10 times bid. So, this information would be very useful in deciding whether a coin was a worthwhile acquisition, at the seller's asking price, or suggest how much to bid for a coin at auction.
When I evaluate a toned coin and try to arrive at what I think is a fair price relative to the market, I mentally use most of the same criterion you have mentioned, but have never attempted to put this method down on paper. I've been wrong many times about the value I have placed on a particular coin, but fortunately, I've been correct most of the time. Values definately change with the market mentality and the current mentality is what you have to be astutely aware of whether you are just a collector or a dealer(maybe more so if you are strictly a dealer).
As far as Andy goes, I can't tell you how many thousands of dollars I've spent on beautifully toned coins that he has offered over the years. I personally think that Andy is one of more astute dealers of the last decade and has tremendous knowledge in the field.
and certainly a challenge.
I do agree, technical grading leaves out, for the most part, what should be the single most important factor in
the valuation of a coin, which is the symphony of all those elements that comprise what we call eye-appeal.
anaconda
Where does this one rate on your scale and what do you think it is worth (1985-O PCGS MS65)?
(Which or course brings up another question, which is can a novice accurately assign eye-appeal based simply on aesthetics or
does one's knowledge that emerald green Morgans are scarce enter into the assesment?)
After reading it a couple of times I realized that I've been buying toned coins using your system!
GSAGUY
P.S. Remind me to give you a discount on your next purchase.
You’re trying to quantify art.
No matter how many categories and how much weight you give each category, it comes down to assigning a numeric value to a subjective opinion.
As far as Anaconda's coin, I'd rate it Dang, that's pretty but I'm not sure it would rate 91 or a 92? How pretty is it?
My posts viewed
since 8/1/6
And, Relayer notes that "it comes down to assigning a numeric value to a subjective opinion."
Under the approach I proposed, no one need assign a numeric value or directly determine eye appeal. In fact, all they have to do is to make fairly simple objective evaluations of certain coin traits. Once the quantitative expressions of these traits are plugged into the formula, the eye appeal measure is then computed for them.
Suppose, for example, that research indicates there are 6 relevant characteristics to eye appeal: predomininat color type, number of different colors, proportion of coin that is toned, location of untoned areas if any, vibrancy of the colors, and underlying luster. Within each category, a measure is selected based on a predeertermined scale and ordering. Suppose all the scales run from 1 (least desirable) to 10 (most desirable).
For some attributes, the approach is simple. E.g., for proportion of coin toned, you record a 10 if the coin is fully toned, a 5.5 if half toned, etc. For color, a preference scale would have to be established. At the extremes, brown black might get 1, emerald green perhaps 10. And so on.
Once these scales were specified and the formula developed, a novice would only have to look at the coin, and simply record objectively those traits according to the scale provided. Then he would plug the values determined for each of the 6 traits into the formula, and the eye appeal index would be easily computed for the novice, without the need for a direct, subjective, and uninformed assessment of eye appeal. In effect, the novice would be relying on the expert's ranking of a representative sample, and that expert's inference of the charateristics of the coins that resulted in that ranking, to find the eye appeal level of a new coin of interest to him, the novice.
Now, to go one step further, suppose PCGS did this research, published the formula, AND THEN made available scans of the sample used to determine the ranking, with the ranking noted for each coin. A novice could then calculate the index for a coin he was interested in evaluating, such as the one shown above by Anaconda. Suppose he determined that the eye appeal index, on a scale of 1 to 100, for this coin was 70, which is about what I think it is compared to my sample. The novice would then know that this coin was in the 70th percentile of coins that were worth between 2 times bid (with index equal to 1) to 16 times bid (for index equal to 100. But, with the scans and ranking data available on the PCGS website, he could then look up which coins in the sample were ranked around 70, and have some idea of the look associated with coins in that eye appeal range, as well as having an sense of how much such a coin is worth.
Hard to believe that no one has actually atempted to do this. Does anyone need a dissertation topic for obtaining a Ph.D. in Numismatics? Wouldn't it be nice if PCGS or NGC would be willing to fund such research?
While I do think that you can take a world class toned Morgan and get 90% of knowledgable collectors to say that particular coin would be a 9+ out of 10, there is just too much personal taste involved in toned coins (color, pattern, location, etc) to ever devise a reliable grading scale that most collectors would conform to IMO.
dragon