Posted 02/17/2021 04:57 am
As I understand it, the Numista Rarity Index is exactly accurate (modulo some low-level timing troubles) in terms of how rare the type in question is among Numista member collections. Numista audience skews French, so French-associated coins tend to have lower ratings and coins that French people are less likely to collect tend to have higher ratings.
This also conflates rarity with interest in general; types that aren't, technically speaking, rare but also aren't collected often (by English/French speakers, at least) would get high rarity ratings.
In addition, the ratings are intrinsically type-based, so a common type coin has a low rating even if the type is full of rare dates, and type granularity has a big impact on the rating (e.g. many 18th and 19th century Russian coins are subdivided into "types" by mint, which makes their rarity ratings artificially higher; for some other series, particularly among ancients, the situation is even worse).
There are probably also many cases (especially among ancients) where the relevant types just hadn't been listed yet back when most members were entering their collections. There's a long-ongoing project on listing all the Roman types (currently up to Macrinus, 218 AD), and after a few years we'd probably see the Roman coin rarity ratings stabilize a little as comparatively common types get the appropriate amounts of collectors entering them (though most types will still be 100/97/95, obviously).
This also conflates rarity with interest in general; types that aren't, technically speaking, rare but also aren't collected often (by English/French speakers, at least) would get high rarity ratings.
In addition, the ratings are intrinsically type-based, so a common type coin has a low rating even if the type is full of rare dates, and type granularity has a big impact on the rating (e.g. many 18th and 19th century Russian coins are subdivided into "types" by mint, which makes their rarity ratings artificially higher; for some other series, particularly among ancients, the situation is even worse).
There are probably also many cases (especially among ancients) where the relevant types just hadn't been listed yet back when most members were entering their collections. There's a long-ongoing project on listing all the Roman types (currently up to Macrinus, 218 AD), and after a few years we'd probably see the Roman coin rarity ratings stabilize a little as comparatively common types get the appropriate amounts of collectors entering them (though most types will still be 100/97/95, obviously).
