The problem I pointed out in the first place is the turning of one way of expressing a number into the only way with any other variation being wrong. The Roman mint produced these two Titus coins from TRP 9. Sorry my elephant is worn but it reads TRP IX later in the same TRP year (1 July 79 - 30 June 80, I believe) but after the IMP XV and COS VIII (1 Jan 80) numbers had incremented. Why the change? IDK.
I agree with Doug's point, that rules referred to in other posts are modern efforts to standardise usage of Roman numerals, which were not followed by the Romans themselves. Here's a pic I took two weeks ago of the Fasti Praenestini from Rome's Palazzo Massimo - part of an early calendar by Verrius Flaccus, the tutor of Gaius & Lucius who would presumably have known his Roman numerals - http://penelope.uchicago.edu/~grout/encyclopaedia_romana/calendar/praenestini.html 18 is consistently written here as XIIX - e.g. see the rightmost column. The subtraction rule isn't used here for numbers less than 18 - i.e. 4 = IIII, 8 = VIII, 9 = VIIII, 14 = XIIII; but is used for 18 & 19 - XIIX, XIX. Sorry - that was unreadable - here's a zoomed-in version of part of the rightmost column/month: ATB, Aidan.