I hold a Masters in EE (control theory). I have never really found studying Mathematics formidably intimidating, probably because I've been at it for some time now.
But I totally suck at the following things:
- Solving problems that require really inspired "crux" moves. Think IMO level problems, lighter versions of Euler's solution to the Basel problem (http://en.wikipedia.org/wiki/Basel_problem) etc.
- Coming up with abstractions and new Mathematics. Think (again) Euler's invention of graph theory or Cantor's invention of Set theory.
- Recalling what I've learned just months after their immediate use is over. For example, I had to cover quite a bit of math for my BEng. But if you ask me now about real analysis, PDEs, stochastic processes or fourier transforms, I'd be stumped.
I strongly suspect that this is because we haven't analyzed and understood the problem of learning Mathematics well. To give you an example, 200 years ago people were trying to get fit. But it was not until the invention of systematic strength training with free weights and machines, etc., that we could really do a reasonably good job of it efficiently. Early body builders look like sticks compared to monsters of today.
I am hoping that such a revolution happens in education as well. Before we learn, we must really learn how to learn. Our knowledge of that is very limited now.