> We have pretty good documentation on the performance of currencies over the past 500-2000 years.
Not “Pretty good” enough to detect hyperinflation by the definition at issue (50%+ monthly inflation.) Before regular, systematic, and frequent collection of data for price indexes, we have good enough data to make wide-range estimates of inflation over intervals of years to decades, in the best cases, but hyperinflation (whether or not it produces collapse of currency) doesn't continue long enough to show up in that, and also hyperinflation tends to correlate with conditions which reduce the availability and survivability of data. (The author of the article here notes a data problem with Yugoslavia in the late 20th century, for instance.)
Not “Pretty good” enough to detect hyperinflation by the definition at issue (50%+ monthly inflation.) Before regular, systematic, and frequent collection of data for price indexes, we have good enough data to make wide-range estimates of inflation over intervals of years to decades, in the best cases, but hyperinflation (whether or not it produces collapse of currency) doesn't continue long enough to show up in that, and also hyperinflation tends to correlate with conditions which reduce the availability and survivability of data. (The author of the article here notes a data problem with Yugoslavia in the late 20th century, for instance.)