Featured 1797 S-139 Large Cent "Suspect" Examples and Initial Research

You would have to calibrate the software with super high grade examples of each variety, and then AI to determine what are just bumps and scratches and what are die diagnostics (devices, die cracks, die scratches, strike weakness, etc.). This is exceptionally difficult for lower-grade/heavily-damaged coins in which the diagnostics can be greatly distorted. However, that much alone is not good enough for these fakes. The repeating marks need to be seen on multiple examples for it to be flagged as counterfeit. So you would need to scan every single coin to get a full profile. You would also need to be able to differentiate the strike characteristics of the counterfeit from damage like whizzing or smoothing.

Then there is the problem of differentiating depressions from worn-down contact marks. Even under magnification, they look the same, hence the need for multiple examples.

 

Log in or Sign up to hide this ad.

The software was developed almost 30 years ago. PCGS had software that would develop a "digital fingerprint" of each coin. With the really good counterfeits, they are detected due to repeating defects which were most likely in the source coins and are now in the die and which are repeated on each fake. Today they are discovered by people who happen to remember having seen them before and if they happen to have more than one of the fakes there side by side. The software would be used to scan coins and eventually these repeating defects would be noted. It would have a "better memory" and could do comparisons much faster.

 

In looking at Hawaiian Copper in Heritage Archives, I've noticed a drastic improvement in the image quality starting about 2006. Right now I'm in the mixed image quality era.

 

Right and it would instantly tell if there was an incorrect die marriage or some other incorrect die pairing. But it would have to be proprietary because if the counterfeiters ever got ahold of it, the fakes would rise to a whole new level.

 

And why is that? They don't need to get a hold of anything. It seems to me that they can strike two fakes from the same die and do everything the TPGS does.

 

Because the counterfeiters could image their fakes, run the images through the software and it would tell them which markers are wrong for the series type, date and MM and tell them specifically which markers need to be "fixed" without having to have very much experience in the series. Some people think just posting the "tell" markers on sites like this one helps the fakers to refine the counterfeits, having software that has that info built in could be catastrophic to the hobby/business.

 

The counterfeiters of these are aware of the "tells" or matching circulation marks prior to us posting anything and have in two examples created a 2nd "variety" of the fake by trying to change/ improve them. We try to get the word out more to the Hobby as an aid to fighting these; there are physical details we leave out of forums like this or the summary articles...

 

I'm definitely NOT in the "IGNORANCE IS BLISS" camp. I read about that somewhere a long time go.

 

Counterfeit detection for a coin struck from a pair of dies made from an authentic source coin has 3 levels of die markers subject to authentication.

Level 1 are the die markers of the authentic die pair used for striking real coins. A real coin should possess all the of these die markers. A coin that does not exhibit all the die markers of an authentic coin would be a suspect counterfeit. As these would be die markers present on the source coin that were not transferred to the counterfeit dies.

Level 2 are the elements unique only to the source coin which are transferred to the counterfeit dies. These markers would be common to the source coin and all counterfeit coins struck by the counterfeit dies but would not be exhibited by any other authentic coin except the source coin. Any coin, except the source coin, exhibiting these markers would be a suspect counterfeit.

Level 3 are the counterfeit die markers. These are produced during the actual manufacture of the counterfeit die. These markers are unique to the counterfeit dies and will be common to all counterfeit coins struck by the counterfeit dies. These markers would not be exhibited by any authentic coin, including the source coin, because authentic coins were not struck by the counterfeit dies. Any coin exhibiting these markers would be a suspect counterfeit.

While a counterfeiter can always produce new and better dies from a level 1 and level 2 marker aspect; level 3 is a different issue altogether. The newly counterfeit dies will contain their own set of unique markers created during their manufacture.

 

I have personally stopped buying them myself a few years back unless I see absolute PROOF someone has had it in their personal possession a few years. Otherwise I won't buy it.

 

Provenance has always been important for me; knowing when these latest counterfeits "surfaced" is also a key to purchasing an example I am interested in...

 

Unfortunately, provenance has only become important to me since late 2019.

I have a note in paypal that I made a private purchase from Donald Stoebner before discovering his 1793 S-16 was an NC-6. Small world. But now I don't know which coin I purchased from him. I just placed coins in their cotton lined envelopes and threw away the provenance.

Live and learn.

 

The Coin Week article just broke 10,000 reads- not bad for trying to get the word out!

 

Just an image final summary; same coin, 3 different results!

 

are you kidding me? same coin? WOW!!