For all series from E48 through E192, this holds true:
$$ value = 10^{\frac i n} $$
where i is the zero-based index in the series and n is the length of the series. The one exception is E192[185]:
$$ listed = 9.20 $$
$$ expected = \text{round}\left( 10^{\frac {185} {192}} ,2\right) = \text{round}\left( 9.19479... ,2\right) \approx 9.19 $$
It seems that industrially, 9.20 is indeed much more common, but to muddy the waters even further, Vishay offers both of these parts:
RN55E9191BRE6RN55E9201BRE6
What is going on here?
There is a random post on EDAboard that talks about the same issue.
This is related to, but not the same as, A question about E12 series resistors - it's well-known that E24 and below differ from the exact logarithmic formula for historical reasons, but no such historical reasons are listed in (for instance) the Wikipedia article for the higher series.
