This question shows that Roman numerals did not take advantage of the regularities that were present in Latin. Quīnquāgintā (L) is clearly built from quīnque (V), yet their Roman numerals aren't related. The numeral for octōgintā (LXXX) is only structurally similar to octō (VIII). I wonder why they didn't have a symbol representing the morpheme gintā (say, G), and write, for example, "VG" for quīnquāgintā and "VIIIG" for octōgintā.
Can this be an indication that the Roman numeric system was created by someone who spoke a language different from Latin? Or is it common that the numeric notation is less regular than the actual pronunciation of the numbers? In the case of Chinese, the number notation seems to match pretty clearly with the spoken language. Which case is more common, and do we have any explanations for the cases where the two don't match?
Update: The answer given so far indicate that incomplete match between Roman numbers and Latin is not particularly strange. Note that I'm asking about the wider picture: is it more common for the systems of writing number to match the language (as in the case of Chinese), or do they usually evolve independently?