Registry Set Statistics Revisted
Gerry
Posts: 456 ✭
PCGS has recently added two columns of their population numbers to the registry listings - the number in the grade of the coin and the number in higher grade.
These numbers may be of interest, but as raw numbers they do not greatly facilitate one’s understanding of how well a set grades, nor do they allow convenient comparison of the “quality” of the various coins in a set. That is, a 66 in one issue may be a lot different than a 66 in another.
Let me suggest, in this message, an alternate way of using those two columns that I respectfully suggest will be more interesting and maybe even more meaningful in reviewing the content a set.
These two statistics are just different ways of looking at the pop numbers. Unless you agree that the pop numbers, which everyone knows are subject to some error, have SOME value then you wouldn’t like these statistics any more than the pop numbers themselves. But let’s face it guys, PCGS grades are what you’ve got, and PCGS naturally wants to base ITS registry on ITS grading system.
The first statistic is what I call RANK: if the coin is in the finest known category it gets a rank of 1, if it has a grade one below finest known it’s rank is 2, etc. Some registry sets, in fact, like to indicate for some of their coins whether it is “finest known” or one grade below “finest known.” So showing ranks in a column next to grades in the registry is simply an extension and formalization of that concept.
If you have a listing of the coins in a registry set with the ranks listed, and you just glance at that column, I submit that you can get some idea of the quality of the set in terms of the appearance of 1’s, 2’s, 3’s, etc. – more so than just looking at the raw pop numbers.
The second statistic is what I call P-TILE (short for percentile –everyone that ever took a SAT exam must remember that!). It indicates the overall position of a coin in its relevant population. Let me illustrate the calculation, as I believe it should be done. Suppose you had a MS65RD 1914D Lincoln – 170 of these coins have been graded (in RD), 124 are below 65, 43 are in 65 and 3 are higher than 65. I would figure that your 65 “is better than” the 124 lesser graded coins and (important assumption) better than one-half of the 42 other equally graded coins. Taking this as a percentage of the other 169 coins in the population, your MS65RD gets a P-TILE of 85.8%. A MS66RD would earn you a P-TILE of 99.4%!
Now if you had an MS65RD 1939 Lincoln, it would get only a paltry P-TILE of 3.0%. A logical difference, right? A couple of other observations: Note that pop 1’s automatically get a rating of 100.0%. Also, these ratings change, for better or worse, as the pop for a coin changes! Hope that doesn’t bother you – I believe it is the way it should be.
I respectfully suggest that you do the proposed calculations of RANK and P-TILE for a few of your coins – I think you will find it interesting and revealing.
Finally, let me say all of this is to encourage PCGS to provide more interesting statistics for a registry set. It doesn’t attempt to remedy what some have suggested are other deficiencies with the registry concept. I’ll leave that to others.
These numbers may be of interest, but as raw numbers they do not greatly facilitate one’s understanding of how well a set grades, nor do they allow convenient comparison of the “quality” of the various coins in a set. That is, a 66 in one issue may be a lot different than a 66 in another.
Let me suggest, in this message, an alternate way of using those two columns that I respectfully suggest will be more interesting and maybe even more meaningful in reviewing the content a set.
These two statistics are just different ways of looking at the pop numbers. Unless you agree that the pop numbers, which everyone knows are subject to some error, have SOME value then you wouldn’t like these statistics any more than the pop numbers themselves. But let’s face it guys, PCGS grades are what you’ve got, and PCGS naturally wants to base ITS registry on ITS grading system.
The first statistic is what I call RANK: if the coin is in the finest known category it gets a rank of 1, if it has a grade one below finest known it’s rank is 2, etc. Some registry sets, in fact, like to indicate for some of their coins whether it is “finest known” or one grade below “finest known.” So showing ranks in a column next to grades in the registry is simply an extension and formalization of that concept.
If you have a listing of the coins in a registry set with the ranks listed, and you just glance at that column, I submit that you can get some idea of the quality of the set in terms of the appearance of 1’s, 2’s, 3’s, etc. – more so than just looking at the raw pop numbers.
The second statistic is what I call P-TILE (short for percentile –everyone that ever took a SAT exam must remember that!). It indicates the overall position of a coin in its relevant population. Let me illustrate the calculation, as I believe it should be done. Suppose you had a MS65RD 1914D Lincoln – 170 of these coins have been graded (in RD), 124 are below 65, 43 are in 65 and 3 are higher than 65. I would figure that your 65 “is better than” the 124 lesser graded coins and (important assumption) better than one-half of the 42 other equally graded coins. Taking this as a percentage of the other 169 coins in the population, your MS65RD gets a P-TILE of 85.8%. A MS66RD would earn you a P-TILE of 99.4%!
Now if you had an MS65RD 1939 Lincoln, it would get only a paltry P-TILE of 3.0%. A logical difference, right? A couple of other observations: Note that pop 1’s automatically get a rating of 100.0%. Also, these ratings change, for better or worse, as the pop for a coin changes! Hope that doesn’t bother you – I believe it is the way it should be.
I respectfully suggest that you do the proposed calculations of RANK and P-TILE for a few of your coins – I think you will find it interesting and revealing.
Finally, let me say all of this is to encourage PCGS to provide more interesting statistics for a registry set. It doesn’t attempt to remedy what some have suggested are other deficiencies with the registry concept. I’ll leave that to others.
0
Comments
One other thing that is provided in a backwards way is a population report. If you collect a series and are thinking of buying a specific coin, all you need to do is find that coin in someones registry set and you will have the population. Great feature PCGS!!!
So for these reasons I say lets leave well enough alone here.
1. Relying upon pops in a pop report is tricky business. For example, by best estimates I have heard, approx. (10) 1964 SMS quarters were struck at the mint. Yet, no less than 22 have already been graded by PCGS-probably just pure regrades with tags not being resubmitted. Yet, relying upon a pop report here would be a bad idea (and yes, I know the 64 SMS quarter is not yet a required Registy coin, however, it is the point I am illustrating here).
2. The P-Tile still fails to take into account the individual coins, coin by coin. For example, if, and when, a 1909(svdb) is ever graded in PCGS-MS68RD, as compared to a 1955(s) Lincoln in PCGS-MS68RD-both get nearly the same value under your system? One being potentially a $100,000 coin and the other a $5,000+ coin??
Gerry, I just don't believe there is any reasonable substitute than value weighting each coin in each grade.
... both get nearly the same value under your system? One being potentially a $100,000 coin and the other a $5,000+ coin ... and ... I just don't believe there is any reasonable substitute than value weighting each coin in each grade
C'mon Mitch, you are sooooo close to just saying "market value weighting", let it out!
Weight it just like you want, but put a dollar market value on each grade rather than an obscure number -- we've got enough numbers in this hobby.
However, in practice, I agree you can't let one or two sales rule the day, and in fact the market values (in the PCGS price guide, in my proposal) would be set by an "expert" or team of them, and would reflect a combination of what's happening in the market and common sense. In other words, an "expert" would interpret the marketplace and come up with a reasonable market value.
For example, the recent PR66 Walker sale where the price paid was very high was apparently because of it's upgrade potential and attractive toning, not because it was a typical PR66. It is also now already apparently in a PR67 holder, and the owner disappointed that it didn't make PR68. So, an expert would not put $22,000 in the PR66 price guide, rather he'd take into account that it was bought as a PR67 shot 68 and perhaps adjust the prices in the PR67 and/or PR68 column.
The beauty of the market weighting system is that dollar values are very understandable by all parties involved, and can be convincingly argued for change when incorrect (for example, as you pointed out with FS Jeffersons lately).
It's much easier to say "this FS Jefferson regularly sells for $2000, not $50 as in the price guide" and have people agree with you, than "this FS Jefferson is very difficult to find, and hence deserves a 4 weight in MS63FS, an 8 weight in MS64FS, and a 172 weight in MS65FS, especially compared to this date which has weight X in grade Y, etc...."
Market value is the shorthand way all collectors and dealers already use to describe a particular coins merits and desirability.
Tad: "Apparently this" and "apparently that"-a proper Registry valuation table isn't formulated on hearsay and gossip!
Wondercoin.
And as you pointed out above in response to Gerry's proposal, they can't rely on pop reports alone for this, because those are notoriously inaccurate in certain important cases. Now there's a hearsay problem.
Another problem (I'll win you over yet!
On the other hand, I think you'd have many experts eager to contribute to a more accurate price guide.
Why can't the registry be regarded for what it is, and not what it isn't.
As for the Walker, it "is" in a 67 holder, and may get in a 68 holder. If it does not, then the buyer may lose on it. I personally looked at the Eliasberg 1885-O in a DMPL holder. The dealer who bought this paid well over $10,000 for this coin hoping it was a Branch Mint proof. He ended up selling this to Michael Casper (who had it in the FUN sale) for a substantial loss. It did not sell there, although it was bid higher than the price paid (I know this for fact!)
Some of these sales have to do with hopes, not anyting registry related! Some people actually still deal with coins, not the registry.
Perhaps the people who are "modern" experts should stay out of the classic areas.
Personally, I liked the older simpler registry, no weighting at all, keep it a game with simple rules, especially since no individual coin's desirability can be reduced to a grade on the label. However, I'm clearly outvoted on that
The market values used for weighting would be taken right out of the PCGS price guides, and as such collectors and dealers interested in registry sets would have a vested interest in keeping an eagle-eye on that guide, keeping it accurate and current.
This would actually provide a valuable service to other collectors who don't care about the registry -- by helping maintain what could be the first truly realistic and comprehensive price guide.
Coins such as that Franklin would clearly be one of the more difficult coins to come up with a representative price guide value, but as of today (and not being a Franklin expert) I'd say $35,000 looks like the best approximation... there are at least 3 people in the world that were willing to pay that much.
My original suggestion was simply to report population data in a certain way and I stopped short of suggesting that those statistics be used to evaluate the sets themselves. Certainly, population numbers, with all their blemishes, have SOME value. I can't resist this aside to my good pal: Mitch, how many times have you quoted the pop numbers when I've bought coins from you?. In any case, I was just suggesting different ways of reporting those pop numbers.
Mitch and Tad, however, get into a deeper more significant issue: overall evaluation of a set; and as I read their comments, I think they are in basic agreement. Ideally a system would give each coin IN EACH GRADE a different weight or value. The current PCGS system only gives each issue a weight which is applied to the grade, and that weight does not vary by grade.
The only difficulty with what SC and WC propose is not in the concept, it is in the execution of that concept. Poor David Hall is sitting back in his office with the responsibility of DOING something, and not just thinking and talking about it. Where does PCGS obtain these values or weights by issue and grade? He needs an operational procedure (and hopefully one that is not too costly) in all its gory detail - whether it be by a survey of experts or by analysis of auction results or by whatever.
I don't know if it adds anything, but since my background is really that of a baseball card collector, I'll mention that price guides for cards exist galore. In fact, every month I receive a publication called Sports Market Report (which by pure coincidence is a Collectors Universe Company publication) that lists hundreds of vintage sports cards and prices by grade. It appears to me that it is much more elaborate and kept much more up to date than the price guide PCGS keeps online, or any price guide I have seen for coins.
So, PSA seems, at least for cards, to have those values that Tad is seeking, and in doing so I'm sure they have solved the problems that Mitch raises. (Mitch, believe me when I tell you that I can quote you variations and crazy prices for cards that are even more unbelievable and wilder than the ones you throw out for coins.)
In any case, is this PSA effort a model of what you all would like to see for coins? Please, please, everybody, whatever you do, don't give me that old line: "but our problem is much more complicated than theirs."
Gerry: My point exactly. Even the "far advanced" PSA system really doesn't work right for Registry purposes
And, Tad's point that the Pop Report is hearsay also, IS MY POINT AS WELL!! When did I ever say to rely solely on a pop report? It can be as foolish hearsay as the motivations of bidders on a proof Walker, if not simply used as one piece in the puzzle of a "value-weight" of a coin.
And with all due respect, dbddie55's comment about modern guys sticking to moderns has absolutely nothing to do with this discussion at all I'm afraid. Value-weighting sets has nothing to do with that unrelated point. But, this modern guy is now one of the strongest buyers in the country for finest known Mint State Liberty and Shield Nickels. Anybody want to sell this "country boy" a few coins from an era in time where SBA actually lived? Only catch is they need to be "4 and 5 figure" coins. I'm crossing the line in a BIG way
One huge advantage of basing a weighting system on your P-TILE rating would be that while imperfect, once it was set up it would require no further maintenance or involvement by experts -- ratings would automatically adjust based on changing pops. As you point out, this probably has great appeal to David Hall, et al, and may be worth proposing to them.
Since all the weighting systems in the world ignore the basic fatal flaw pointed out earlier by dragon, I think your system has arguable merit.
In other words, putting blind trust in PCGS pop reports to determine rarity is certainly no more worse than putting blind trust in PCGS grades to determine the "finest" set.
But, I'm done arguing your case, back to mine...
Mitch, I realize you didn't say to rely on pop reports -- that was simply a counterpoint to your argument that we shouldn't rely on hearsay to weed out artificially high auction results (i.e. the Walker proof was undergraded). My point is that hearsay is also required to figure out the true rarity in your proposal (i.e. the pop report is invalid), which is at least as difficult. Things like the Walker sale tend to be more public. Somebody wrecking the pop report by resubmitting a 64 SMS coin 10 times without turning in their old labels isn't exactly going to end up on the front page of Coin World.
Really, I don't think I have heard any argument against market-value weighting that doesn't also apply similarly against weighting every date/grade combination.
And market-value weighting has distinct advantages as I've outlined -- to registry and non-registry collectors alike.
Tad: Then you haven't been paying attention
Market valuing weighting is what has the 1953(s) nickel currently at $300.00 in MS66 grade with my sight-seen buy price of $15,000.00!! You think you can simply correct that price? How, based upon my "stupid" crazy high open offer, or my low-ball offer that might make me some good money because I know the real value of that coin. Which is it? So, you want to "play it conservative" and raise the $300 to $5,000? Or, just follow the "bouncing ball" and raise it all the way to my offer? Or, ignore my "silly" offer and keep it $300? Or, make it $20,000, because you know Wondercoin needs to make some profit on his buy price? What do you base this decision on? Years of study on that date in that grade (like you know your Ikes for example) or "following the bouncing ball"? Employing an expert in the field of MS Jeffersons to properly value weight the series is the ONLY logical solution. Gerry, I believe you know it too (you are just a bit upset that NGC won't reveal its weights-right?) It costs money to properly value-weight a series, which NGC paid to have these series value-weighted for their registry the right way.
Tad-do you why you are arguing a losing cause here? I'll tell you imho. The #1 expert in the country on Lincolns for example MIGHT agree to provide the true value-weights if the grading company, like NGC, intended to keep the chart confidential (as they have). But, to give away 20 or 30 or 40 years of valuable information to correct every price guide-these dealer/experts make their livings working from erroneous price guides-right? I'm afraid you are in "fantasyland" here
Mitch: what do you see as the basic difference between market-value weighting every date/grade combination and your idea of weighting every date/grade combination? Doesn't it just come down to how that value for each combination is determined?
What is needed is an operational definition by which those values or weights are determined. In other words, how are those values determined? I would have no trouble with a panel of experts. Any volunteers?
Rating by market value has two problems that bother me. First, the coin market (especially at the high-end) is too volatile. Today's "finest" set will fall tomorrow as relative values (even within the same series) change. Rating by market value will not end debate about who really has the best set.
Second, market value does not always equate to difficulty in obtaining coins. Some coins that are extremely rare (and very difficult to find)have lower market values than other coins that are less rare (and easier to obtain) because of differences in demand.
I think set rating should reflect a balance between market value and rarity/difficulty. I'm afraid I don't know how to do this. However, I think a simple weighting approach is more accurate than non-weighted scoring, and probably better than a volatile market value rating approach.
I disagree that dealers don't want an accurate price guide, and in any case there are plenty of other inaccurate guides that unethical dealers could use if they wanted to take advantage of someone.
Even if there are those expert dealers who won't share pricing info, there are probably more that would gladly share pricing info but would not care to muddle their heads with a bunch of mystical rating numbers.
And, again, the beauty of the pricing guide is that it IS in the open, that's a strength, not a weakness! It's there for everyone to see, and make meaningful arguments for or against any price that is out of whack.
Tad: Fantasyland again
1. If dealers want an accurate price guide, would you mind telling why it hasn't happened in all these years? Why are Lincoln prices (a set which has been popular for years) on the guide terribly off after all these years? Why haven't the KNOWLEDGEABLE dealers rushed in to correct the guide. In fact, why haven't the KEY COLLECTORS rushed in the correct the guide? Why? Don't waste your time writing back on this question-I'll answer it. Because those key collectors like it just as much as the key dealers. How do you think a collector feels if he buys at the show a 1953(p) nickel in PCGS-MS64FS from a dealer for around $50 because the guide says $40 and then, hypothetically, proceeds to sell it for $4000+. Expect that collector to call CU the next day to ask them to raise the price on the guide?
2. "Unethical dealers" taking "advantage of someone". "Unethical (highly educated in the series) collectors" ripping off unsuspecting dealers who look at the guide for reference- This is all simply silly talk Tad. Simply "fighting words" designed to rally up the "newbies" to your position. With all your years in the business as a dealer, you should know better than to write silly stuff like that
Just my humble opinion
As far as the fantasyland I'm living in (Ikeland), I know of at least one prominent Ike dealer (not me) and at least one serious Ike collector (also not me) that have invested significant amounts of time petitioning publications like Coin World specifically to get them to raise the "trends" prices of high-grade Ikes to more realistic levels. And neither person, to my knowledge, is (a) stupid, (b) philanthropic, or (c) sadomasochistic.
But even assuming your conspiracy theory is true, that all the experts/serious collectors in a field want to keep the price guide inaccurate for whatever reason... do you think that would still be the case if PCGS decided to weight the registry based on price guides?
At that point, you've now got an army of registry collectors -- in many cases the best customers of the experts in the field who supposedly want to keep things hidden -- that need the price guide to be accurate to get credit for their tough coins.
Do you really think those collectors would still desire the price guide to be too low? And do you really think the experts putting together the guide (likely dealers) would ignore their best customers' requests for accuracy? Want me to throw a silly fantasyland remark in here somewhere?
Tad: You're not suggesting the Grey Sheet pricing on gem Lincolns is accurate are you? Want to ask Gerry to comment on that one ") Wondercoin