lucy24 - 2:26 am on Feb 20, 2013 (gmt 0)
I meant google. At any given time in any given search, there will be some legitimate redirects. It's a byproduct of caching: the version you saved may not be the version that's there now. So if you're serving up content to users-- as in "view original image" --you can't ignore all redirects all the time. Otherwise it would make the search engine look inept. "We're just got through saying this image exists, but now we can't find it."
But if it's an ordinary redirect, the new image would be expected to fit into the same visual slot as the old one.
If the page says <img src blahblah width = "abc" height = "def"> and your cached copy has physical dimensions abc and def, then you might have grounds for suspicion if you are redirected to an image with different dimensions. Filesize wouldn't matter, because that's not reflected in the originating html.
Topic for further investigation*: Does anything behave differently if the original html did not specify size attributes in pixels?
* By someone else, because all my <img> tags do give explicit pixel size.