GG explains the differents between feed-based and crawl-based Froogle results.
[Link to quote
pmac/mona, if you take a search and start adding more and more specific words, at some point you'll see the breakdown. Um, here's a search:
That page includes results from feeds where we can reliably tell the price. Further down are results from crawling the web where we're not certain about the price, but the results may still be useful to searchers. See also
(question 7). Look for a green bar that says "The results below were automatically extracted from web pages. Price and category information are uncertain." We can't be positive that those prices/snippets/images are correct because we found them as we crawled the web, but there's a good likelihood that users would find that info helpful, so we include it after the data that we're certain about. It helps a lot when looking for obscure stuff, for example.