Oh, hey! There's a good idea!
I just did something similar for a pinball design site whose search engine was completely out of it.
So you're saying we slurp the data via repeated URL calls (perl works a lot better for that, by the way) analyze a couple of resultant HTML files, figure out where the relevant data starts, only publish sub-files from that line onwards (you need to run scripts like that the same day you find that out), analyze the sub-files and publish our own MySQL table based on the info stripped out of those file fragments? (PHP is a lot better for that, by the way)
I think we could do that. Where do we start?