Agreed with other posters that the flaws for the number related stimuli tend to be centered on percentage to part fallacies. But i did notice that these types of stimuli do not tend to show up as much in the more recent PTs.
The ones in the fifties seem to be a lot more abstract and dependent on seemingly little details. I think you just really have to talk yourself through the answers when you're reviewing and start to notice subtle little patterns.