Please venture your opinion, every one can help

Comments

• morganbarber Posts: 1,821 ✭✭✭

  February 27, 2006 4:04AM

  I think it is impossible. If you interviewed every person on the planet, and waited while they went through the dates of every coin in their house, there would still be hoards owned by paranoids who would never tell the truth about what coins they hold.

  I collect circulated U.S. silver

  0
• RYK Posts: 35,800 ✭✭✭✭✭

  February 27, 2006 4:10AM

  Excellent question/problem, and there is no good answer. One starting point would be to consult individuals who have done this in the past (ie. Doug Winter has done this for branch mint gold and Liberty $20's) and find out what techniques they have used. Perhaps you can modify these to suit your areas of interests.
  
  My gut feeling is that your mission will be easier to accomplish in coin series where most of the survivors are certfied by PCGS and NGC (ie. Carson City $20's) than in a series in which most survivors are not certified (ie. mercury dimes). Similarly, the number of survivors are probably more easily counted in higher grade coins, more expensive coins, and coins in which the number of survivors is known to be low (ie. low mintage). As an example, I bet that any one of us could with a weekend of work and the appropriate resources available get reasonably close to the pop numbers for Stellas. Similarly, I could have all of the resources in the world, and I doubt that I could get anywhere close to the number of extant 1887 dimes or 1916 dimes.
  
  I fear that I may be on a fool's errand that may not help me get any closer to real populations than the "feels" and "guesses" currently out there.
  
  The thought crossed my mind. Good luck and keep us posted.

  My Box of 20: https://pcgs.com/setregistry/collectors-showcase/box-20/ryks-gold-box-20/17132

  0
• Longacre Posts: 16,717 ✭✭✭

  February 27, 2006 4:13AM

  This is a great project. As RYK suggested, you should speak with someone who has done this sort of research before. I always wondered how people determined that certain numbers of coins remained in existance. I suppose with ultra rarities, it is easier to trace the populations. However, with coins that are stated as having approximately 500 to 1,000 in existance, I wonder how those sorts of determinations are made. Keep us posted.

  Always took candy from strangers
  Didn't wanna get me no trade
  Never want to be like papa
  Working for the boss every night and day
  --"Happy", by the Rolling Stones (1972)

  0
• MrBreeze Posts: 1,036 ✭✭✭

  February 27, 2006 8:46AM

  bump

  https://www.pcgs.com/SetRegistry/collectors-showcase/world-coins/one-coin-per-year-1600-2017/2422

  0
• MrEureka Posts: 24,513 ✭✭✭✭✭

  February 27, 2006 8:52AM

  Just keep on accumulating data and you will get closer to your goal. You will never completely finish the job, of course.

  Andy Lustig
  
  Doggedly collecting coins of the Central American Republic.
  
  Visit the Society of US Pattern Collectors at USPatterns.com.

  0
• MarineEAS1999 Posts: 327

  February 27, 2006 9:05AM

  how about stamping every coin with a serial number?
  
  no, really...some sort of id number given by the grader even if its an optical scan of some sort. but they would have to charge to do it and i doubt people would want to pay for it. this is a pipe dream at best because of the reasons others have listed already. it would have to have been minted that way but even the mint would have no reason to care about tracking coinage...they mint millions upon millions and it wouldn't be cost effective. truthfully the only ones interested in populations are coin dealers and coin collectors...tpc's may have some interest, but minimal at best. the only coins that populations are important for are the low pop ones...and those may be the ones you can actually get some idea of with a tracking system.
  
  to be certain...if i had a few rare coins of the same issue...i wouldn't tell anyone and i wouldn't try to "move" them at one time...i would manipulate the population to my advantage. that's the only way to make "sweet coin"

  Ken
  
  My first post...updated with pics
  
  I collect mostly moderns and I'm currently working on a US type set.
  
  

  0
• morgannut2 Posts: 4,293 ✭

  February 27, 2006 9:57AM

  Statistically you MIGHT approximate a solution to the question, but there is just no way to sample an average group or series of coins to see how many have been multiply certified or recrossed by NGC-PCGS. More importantly there's tons of material never certified lying in someone's attic!! There are just no records (data) to work from!! A possible approach: In any one date-mint mark-- as long as the population is very, very low-- you can assume an upper fixed mathematical boundary. So, say if you look at say a bunch of low population coins where almost every example is known through auction history (1884, 1885 Proof Trade Dollars, etc)-- you can then assume a fixed point ("answer" to your question) over and over for groups of rarer coins (say pop 25 or less). Then determine a typical resubmission rate for rare-expensive coins as a mean with a plus/minus variance ("deviation") to remove those resubmissions---- always knowing there might be a loner never auctioned somewhere. The next problem then is a lower boundary--in other words, coins so common they either never get resubmitted, OR because they're worth less than the holder ( the cheapys are so common, they never got sent in). One way to estimate the lower boundary might be to use GSA Carson City Silver Dollars as a surrogate. You know the mintage of uncirculated coins released by GSA (just ignor those released prior to GSA - or estimate unc. pre -GSA number from an uncommon one like 1885CC). You know the total uncirc. submissions from the services. So now you know the unknown % uncertified you're looking for (released minus known certified = "unknown" uncertified) and know the variation between what gets certified and what doesn't. Finally you can get a number for a few intermediate populations -say some tougher gold date/MM's, some low mintage 19th Century proofs, etc. The idea is basically you end up with a tight range for the group of very rare coins, a really broad range for cheapos using GSA's (or other know groups), and a bunch of intermediate rages for semi-rare examples with a semi-tight range of uncertainty. Then the idea is given any known mintage, you extrapolate an estimate of remaining total population using the total submissions, or any point you're interested in calculating. This ignors whether the relationships is linear or non-linear (curved) but still might give a decent approximation or range of reasonable numbers, depending on the series. The series is important, because ironically even though I suggest you try GSA Morgans, other Morgan populations are totally messed up by meltings, strange anomalous issue dates (1903 O), etc. Still it might be a lot better than the APPARENT GUESSES of total populations remaining given by Q.D. Bowers in his published Morgan Book!! By the way it might also be interesting to calulate the total number of whizzed/altered/damaged/engraved/ruined coins for few key dates like the KEY 1893-S Morgan, using NCS and ANACS damaged numbers -assuming they keep them. It would be fun to know how many 100's of thousands of dollars have been thrown away by people "improving" just a few key dates!!
  

  morgannut2

  0