Crawl errors and broken links, GWT reports "wrong"..

[dreamcon]

Hi Guys
as part of SEO efforts I work to reduce the number of crawl errors on my site.
I go to webmaster tools and download the list of links that are returning a page not found error.
Whats really confusing me is that from the total number of crawl errors that are listed, only a fraction are actually "valid" broken links (after I cross off the inbound links from external sources).

I am surprised to see that google webmaster tools is listing flash parameters (google shouldn't be able to even read flash right ?), and the numbers of hit box and java script parameters that are getting listed as effectively broken links.

(I look up the source page of the crawl error and check the source code and see that the link string is a parameter of some kind, not an actual link)

Now while I am glad to realize that I am not serving my clients bad paths but I am just wondering is anyone else seeing anything similar. (And naturally are there any way s to avoid it?)

And surely having a 1000+ not founds on a large site is not good, so if only 500 of those are real, its a big difference.