i have found i have a tendency to view past results more positively than they really are. in practice much more patience is needed to follow a system whereas checking past results tends to imply a more positive strike rate.
a bit like the anti-cyclical investment theory. when the papers start to tout it, its time to leave.