You can measure the position and momentum of an electron and the results will be valid. The problem is that in quantum mechanics we are interested in an ensemble of measurements. That is we are interested in the results of a large number of measurements on identical systems. The position and momentum properties of a system are described by two different sets of bases or states. The consequence is that if we find the system in some position state A_x, then for the momentum the state could be A_p, B_p, C_p, etc. This causes an ensemble of measurements to have a certain degree of spread in their results. So if we only look at position measurements that yielded A_x, we may get 10% of them to have momentum A_p, 25% B_p, and so on. Likewise, if we only look at measurements that yielded a momentum A_p, we would get some A_x, D_x, F_x results for the position.
The spread in the results, the variance, is described by the Heisenberg Uncertainty Principle. This principle makes no assumptions about the accuracy or precision of a measurement but is simply a consequence of the mathematics of quantum mechanics. So even if we can perfectly measure the position and momentum of an electron, we say that we cannot simultaneously know the two because over a large set of measurements we will end up with a variance in our results.
Another way to think about it is that we cannot gleam the entire information about a system in terms of position and momentum at the same time. Because a position state can correspond to multiple valid momentum states (and vice-versa), then the momentum of our position state is a superposition of momentum states.