Scrapers exist to take data out of HTML, malformed or not, when an API or feed is not available. Yes, ideally HTML is written in a semantic, annotated machine-parseable way; that also enables smarter search engines, better accessibility, interoperability and so on. That's one of the main reasons behind the changes to the standard made in HTML5.

A better example would be "hey I have this cool device that will improve your mileage by 100%, but it only works on cars built according to current emission standards...".