Madness

[Silly joke removed]

I'd love to see all your scans! Thank you. At the moment I'm concentrating on the sixpences, so it would be great if I could have access to these images.

Thank you @Nonmortuus. Splendid coins.

I don't know if there's such a thing as an undergrad degree in numismatics, but if there were it should probably include subjects drawn from the following disciplines: Social history Economic history Archaeology Probability/statistics Graphical/imaging information technology Curation/conservation Materials science Historical survey of mechanical engineering At least one non-native language A practical subject in the art and craft of coin-making, including experimental archaeology The trouble with such a course is that it would only have niche appeal and little application to a career outside of academia. If only such a degree existed in Australia 25 years ago! Many of my areas of interest seem to intersect in numismatics. My year twelve subjects consisted of Modern History, Mathematics (calculus and functions), German, Physics, English and Music. I guess these are roughly equivalent to A levels in the UK. I took a year of Classical Greek and two years of Koine. I gained entry into an honours degree in engineering, but deferred, never to return. I began a masters in library and information management, again deferring after only one semester. I'm certainly not trying to build myself up here by reciting this sad history of not seeing things through. Rather, I'm analysing the particulars of why I'm drawn into this strangely appealing hobby. Perhaps others were sucked in for similar reasons. Anyone?

Hypothesis 6 - Once new varieties are discovered and brought to the public's attention collectors and sellers will examine their coins for these varieties and themselves make the existence of these specimens known publicly. It follows that this will effect the results of a die study and any attempt to determine the number of coins struck from particular dies. Since revealing the existence of the Hearts/7-String/Serif-7 variety of the 1787 sixpence on this forum two more have been listed for sale on eBay. I guess this could be a coincidence, but the population of this coin has immediately grown by 200% from 1 to 3! Hmm. Is this a good thing or a bad thing for a die study? If a mathematical model could be developed such occurences could be accounted for. Speculation - I wonder if it would be possible to building into a statisical analysis the effect of variants on the use of a particular die. If a die has a scarce variation it will have relatively high survival rate of coins. The run-of-the-mill dies will produce coins that are circulated into oblivion. The concept of "scarce" itself is also then called into question. Dies with insignificant variations may well become scarcer than those with obvious variations and be more difficult for the completionist to find in a high grade. Oh, a die study is a tricky thing! I wish I knew more about the mathematics of probability! Actually, I need to refine that statement. I wish I knew anything about the mathematics of probability! OK, OK. You've caught me out. I wish I knew anything about mathematics!

That's awesome, thanks @declanwmagee ! Thanks everyone. The more the merrier! I have about 280 at the moment.

Thank you!

@Peckris 2 Do you have any images of an over-dipped coin so that I can see an example of their appearance (dull & pale)? Thanks! Does this apply to all metals/alloys, or just silver?

It would be nice to have something published, but I'm afraid I'm far too "newly-minted" and lacking the patina of experience and general numismatic knowledge. I'm only two months into this hobby/obsession. I think I'd need to approach someone much more knowledgeable than me to review the work before presenting it publicly.

It's true that knowledge is power. It's also true that a little knowledge can be a dangerous thing. I'm not sure which of these two statements is true in relation to my most recent purchase: I would guess that this coin should be graded at aVF at best. I paid 60 GBP for it, despite suspecting that this is a few multiples more than the Spink value for a 1787 sixpence in this condition. Why was I willing to pay this much? 1. It's the only example of a 1787 sixpence with this combination of features that I have seen out of the 200 in my growing database: with-hearts/with-serifs-on-7's/7-strings-on-harp. I have another image taken by the seller from a different angle and taken under different lighting that prominently shows all seven strings. There are possibly other coins made from the same die, but they appear to all have only six strings. This die match has yet to be confirmed. 2. It's the most prominent example of an off-centre struck 1787 sixpence I've come across to date. By the way, is there a technical word in *numismatisian* for an off-centre strike? This isn't an example of madness of the "unresearched variety". Rather, I think of it as an example of madness of the "I've-researched-and-want-it-but-have-probably-paid-too-much" variety. One of the unavoidable dangers of attempting a complete collection of varieties and dies of a particular coin is that the only remaining examples are very hard to find, in poor condition and possibly both. How much is something like this worth? To most people, probably very little. To me, much more. Would anyone care to guess what the market price would be for something like this? Is it crazy to pay over Spink?

Notes to Self Hypothesis 1 - If a coin has evidence of circulation wear and is "lustrous" it will have been dipped Hypothesis 2 - If a coin is identified as a scarce variant shortly after being released for circulation it will be quickly absorbed from circulation by collectors. Hypothesis 3 - If a coin is withdrawn from circulation by collectors shortly after being released for circulation it will be less likely to show wear than one that continued to circulate. Hypothesis 4 - Coins with scarce variations that require magnification to identify will be more likely to continue to circulate than scarce variants identifiable by the naked eye. It follows that, in general, the condition of the former will be inferior to that of the latter. Hypothesis 5 - Coins that were withdrawn from circulation by a collector are more likely to be extant than those that were not. It follows that coins with scarce variations identifiable by the naked eye will be over-represented and skew any attempt to estimate or determine the number of coins produced by a particular die. It is likely that coins with scarce variations identifiable under magnification will be somewhat over-represented and influence any attempted to estimate or determine the number of coins produced by a particular die.

A couple of thoughts have occurred to me: 1. Coins are three dimensional objects, but two-dimensional images capturing them can be hard to interpret due to differences with lighting, camera angle, shadows, reflections etc. Is anyone aware of the application of 3D imaging technology to create a computer model of a coin's surfaces? 2. Has machine-learning technology been applied to die analysis? It should be possible for AI to recognise and contextualise features on coins (such as the position of letters in relation to one another) and compare specimens from a large sample in order to catergorise them into being the products of the same die. Sure, there would be complications like wear and "touch-up" engraving, but it's possible that either AI could learn to account for this, or the AI could do the grunt work while a human interprets the results. Pretty keen to hear your experiences and research.

Two songs about rats. As strange as it sounds, Michael Jackson and Mozart share the experience of having been exploited prodigies.

Down the rabbit hole again! I've started to investigate the dies used to produce the 1787 sixpence with the intention of collecting one example of each die (over time). In the past three days I've built a data base of images from 171 examples, hopefully with many more to come. My target is 500 sample images and I hope this isn't optimistic. Here are a few preliminary observations about the reverse dies. There are four primary families of dies: With semee of hearts in Hanoverian Shield and without serifs on the "7"s in the date With semee of hearts in Hanoverian Shield and with serifs on the "7"s in the date Without semee of hearts in Hanoverian Shield and without serifs on the "7"s in the date Without semee of hearts in Hanoverian Shield and with serifs on the "7"s in the date In the overwhelming majority of instances the Hibernian harp contains 6 strings. However, I've come across the following: A. 5 examples of with-hearts/without-serif-7 coins that have 7 strings (just under 3% of sample to date) B. 3 examples of without-hearts/without-serif-7 coins that have 7 strings (about 1.8% of sample) C. 1 example of with-hearts/with-serif-7 coin that has 7 strings (about 0.6% of sample) All of the examples in group A belong to the same die. At least 2 examples of those in group B are from the same die, while the jury is out on the third example. I strongly suspect that all coins in groups A and B are from the same die. My reasoning? Firstly, the absence of the semee-of-hearts was a mistake recognised and rectified part way through the run. As with the shillings, the hearts were hand-engraved, presumably on existing dies, to correct the error. There are about three die flaws that you can see developing across the coins in groups A and B, becoming more obvious as the die ages: Die fill in two letters and the wearing of the left lower serif in "1" of the date. I'm noticing generally much more die wear on 1787 sixpences than on 1787 shillings. Given that a similar quantity of each denomination was struck, I suspect that there were, perhaps, significantly fewer dies used. This is just a hunch. I'll see if can find any records indicating the number of dies produced as this information is available for the shillings. The number of dies produced needn't be the same as the number of dies used, though. Further updates to come.

A few options: Singing Yaks' Organisation Strident Youthful Occultists Splendid Yodelling Oncologists Salubrious Yam Omelets Slimy Yesterday's Oatmeal Sceptical Yearly Optimists See Your Optometrist

I'm in the early stages of conducting die studies of 1787 shillings and 1787 sixpences. Do you have any examples of these coins? It would be awesome if you could post images of these in this thread, both obverse and reverse. High resolution scans are preferable to photos for this purpose, although I'd be very happy with whatever is most convenient for you. By way of thanks, provided you are interested, I'll let you know where your coin/s fit in to the results.

Note to self: Die study methodology It could be helpful to make an attempt at a preliminary grouping of dies for both shillings and sixpences. This could be accomplished on the obverse by literally joining together the dots, forming a quadrilateral. By comparing the the side-lengths and angles within the quadrilateral there may be sufficient differences to form "families" of dies, even if not individual dies. Using a dot would be preferable to a letter in at least one regard, as it is easier to find the centre of a dot than locate a reference point on a letter that might have been disfigured by wear or fill. Perhaps it could even be possible to do this by joining together only three dots and measuring angles and lengths of a triangle. It would certainly be less time consuming as the measurements and mathematics become simpler. This method would allow the use of a simple grid system to define positions (as opposed to circular position and distance from centre) and wouldn't require the rotation of images to standardise positions. The correlation of three/four angles and three/four distances would be sufficient to define the die family. Consider the use of a vector based graphics tool for this as opposed to GIMP. Once grouped into "families" further differentiation can be made between its members to find individual dies.

I'm beginning to wonder about the possibility that the reverses of the 1787 sixpences were manufactured differently to the shillings. The consensus is that the shilling dies were created from fully-formed punches that already had all details in place, including lettering. I'm observing features of the sixpence reverse that suggest each die was built up from a number of punches with different design elements. I've noticed the following: 1. Alignment of lettering differs in at least a few examples, but I've not examined this closely yet 2. In some examples the 7's of the date have serifs while others don't 3. The rotational position of the circle in the centre of the garter star varies In the sixpence pictured above there are details entirely missing at the top of the uppermost two crowns and yet the letters above them are still visible. This suggests to me that when the die was being created the "crown-punch" (if, in fact, such a thing existed) was struck at a slight angle, resulting in coins with a relatively high relief at the base/corner and details missing at the top/opposite-corner. I'd need to find a. other examples of this die to confirm, and b. compare the precise position of each of the design elements across as many dies as I can find. I'm going to buy this coin. At 59 GBP the price seems OK. By way of comparison Michael Gouby has an EF "no-hearts/serif-7" example for 95 GBP. It would be good to have this coin in my hand and look at it more closely. In addition, the details in GIII's hair and the Hibernian harp are amongst the best I've seen on a 1787 sixpence so far. I've decided not to bid on that I/D sixpence being auctioned by DNW, which means I've got some money to spend on examples of other dies.

Note to self: To fully describe locations on a coin face it's necessary to use both circular position in degrees and distance out from the centre or in from the rim. Sector & segment differentiation and alignment of a figure/letter/number could also be of some use in die analysis.

My guestimate: aEF Generally well struck coin, with the possible exception of portions of segment between 300 degrees and 60 degrees on the reverse, most notably the crowns. This, though, could be a punching issue rather than a striking issue. The of rounding of laurel leaves suggest die wear, not coin wear. Older die also suggested by pitting and rounding of lettering. Flattening of shirt frill suggests coin wear. Reverse die is of the "no-hearts/serifs-on-7's" variety. Possibility of pre-strike adjustment marks on reverse. Could @Rob or someone please critique by observations and grading? Thanks! Some weird, silicon-like marks at 0 degrees on obverse and garter star on reverse. Anyone know what these might be?

Thanks @Rob My grading was influenced by the lack of detail/flattening in the curls of GIII's hair. Do you think this is die fill, not coin wear?

Never mind