There is no relationship between the size of the data files and the size of the PAR files
according to their sourceforge site, it works as follows :
The technology is based on a 'Reed-Solomon Code' implementation that allows for recovery of any 'X' real data-blocks for 'X' parity data-blocks present. (Data-blocks referring to files OR much smaller virtual slices of files).
so you could conceivably recover 5 real data blocks for every 2 parity ones, or so. I havent used reed-solomon error correction in a while, and cant remember how exactly THAT works, but i think/thought it was used on cd's.