Seems like there are some situations such as breaking a coin out for re-submission that would affect the numbers. What are others? Are the TPGs diligent about keeping a handle on this? Thanks
If you're talking about population numbers, they are not really accurate, but a good indicator at least.
They're ball pack figures. Most people don't send the labels back in for the TPG to remove it from their population reports. Any grade that has a big price jump in the next grade you can assume that population will be inflated from crackouts trying to the big jump in value. Probably the only way to really get them back to being very accurate at this point would be to delete the whole database and have everything sent back into be counted. Obviously that isn't a practical solution or anything that would happen though.
What coins do you have in mind? The answer will vary greatly depending upon the answer. For example US versus any other country and the market price.
You can consider the numbers less reliable at the higher end, and when the next grade up represents a major increase in price. People will crack and resubmit the same coin half a dozen times seeking the next grade, each time generating another member in the Population at the grade assigned.
Great question and something I've just taken for granted. I didn't really think about resubmissions and how they affect population numbers.
The market value and collector preference are the primary factors in the completion and accuracy of the population reports. As an example on the accuracy, last I checked the 1794 dollar had 14 mint state listed. In a prior PCGS post, my recollection is the owner ("TDN") of the SP-66 stated eight are accounted for and I believe PCGS Coin Facts estimates 10. I'd say TDN is about the most knowledgeable of this particular coin. On the flip side, NGC and PCGS counts are proportionately much lower than the estimates for Baker 24a, the 1792 Washington half dollar in silver, small eagle reverse. The "Washington Pattern Coinage of Peter Goetz" identified 22 but my recollection is less than 10 have been graded. This coin almost never comes up for sale. PCGS contributor "Roadrunner" is a specialist in liberty seated coinage. My recollection is his posts claim one of the rare dates had a single coin submitted maybe 10 times but only someone who is very familiar with a specific coin will be in a position to know it. I believe most 20th century US coins usually have much greater availability than the population reports indicate, except in the absolute highest grades and sometimes maybe one or a few grades below it. In grades of MS and PR 66, multiple times the current counts.
The numbers are highly accurate as to the number of coins they have assigned that grade to. It bears no relationship to the number of DIFFERENT coins or to the number of coins that are still in their slab bearing that grade.
With a word change: "The numbers are highly accurate as to the number of XXXcoinsXXX slabs they have assigned that grade to. It bears no relationship to the number of DIFFERENT coins or to the number of coins that are still in their slab bearing that grade.'
So it seems there is a sweet spot. For lower cost coins, people won't grade at all. (It just takes the value of the coin and gives it to the TPG.) For high-end coins with sizeable price increments between grades, the coins are resubmitted until the owner gives up trying. In the middle there may just be a range where the owner wants it graded for any number of reasons, and lets it stop there. It might be an interesting exercise to look at the TPG's pictures, pick out some identifying marks, and see how many of the coins are the same one.
Proportionately yes, but in absolute numbers, no. There are many individual date/MM combinations (primarily from US coinage but also world coins) which have a relatively low value with high population counts, in the thousands. The biggest outlier is the South Africa 2008 Mandela 90th BD 5R which has over 200,000 business strikes graded. This was the result of a national mania (probably the closest thing to 17th century Dutch tulip bulbs in coin collecting) but there are many others. For practically all of these coins, there are certainly a noticeable to huge multiple, even in high or very grades.
You might be able to do that on NGC as they have a much more extensive database of pictures, (PCGS has a scrawny one) but you would have to know the serial numbers of each of the coins to bring up their images. Frankly it can be a little easier to go to the heritage archives and look up the coin there and see how many times you can find the same coin in different holders and grades.
You can do that on Coin facts for PCGS just pull up all the pictures for that date/mm. There's a lot of examples especially with plus grades where someone has tried a couple times for the next grade up.
