I'm usually pretty good with predictions and broad strategy.
The thing, though, is that you have to stay engaged. You can't just make a detailed prediction early on and then stick with it, you constantly have to be calibrating it with new info and changing your understanding of the picture (or keeping up with the current data map).
I've found sometimes that my emotions will throw me off... or fear caused by an awareness of possibilities (sort of like a trapeze artist who might suddenly be scared of falling if she starts thinking about the drop and everything that could go wrong, and thus mess up). But usually my rationalization of the situation was accurate in the long-term, people ended up with a pretty close response to what I expected. I'm learning to just trust my insight and not let the short-term fears throw me off.
It operates from a very subjective view of the world (i.e., your own value ruleset), but it's not irrational -- the logic follows from your foundational assumptions, you simply have to communicate the ruleset in order for people to follow you, rather than the T approach where the ruleset can more easily be discerned just by observation of the world.
S and N are the irrational functions.