I did badly
40%-60%. The pattern was clear: When I'd seen a picture before I got it right, but when I hadn't, I tried to analyze the picture, and I was sabotaged by the small size and poor quality. That's why I consider every ratty little Web JPEG suspect until somebody can produce the high-resolution original. I think the test basically measured preexisting knowledge of known hoaxes, rather than an ability to analyze new images.
