Shortly after the news was carrying the story about how " being a little overweight is better for your health", I read a follow-up in the news that once the study was corrected by taking out people who were at lower weights due to illnesses, etc, that the study data clearly showed lower mortality in people who were not overweight. I guess that news wasn't as satisfying to report or remember.
In any case, here's a link to an insurance company article, discussing the relative and absolute risks of obesity. It would be great to have you experts in interpreting studies have a look at it:
This one's a PDF, with lots of charts and graphs: