With 15%/50 of the riding considered battlegrounds and too close to call, this kind of prediction is almost useless, but interesting.
Such skepticism. Next you'll be telling me that Ipsos Reid was wrong and that George Smitherman didn't win last year's race to be mayor.

