I am unsure what you are missing because what you explained is approximately correct.
Minor correction:
Entangled qubits (each of 50% propbaility to be up) can have more possible measurement distributions that the two you mentioned.
Distributions like the following exist.
+/+ 20%
+/- 30%
-/+ 30%
-/- 20%
> if I have several pairs of entangled electrons all with spin of 50% up / 50% down, then I might expect the results of measuring the spin of those pairs to look something like this table:I think you may misunderstand how to get qubits entangled. You would have to pass two qubits through a two-qubit gate to get them entangled. And doing so would leave you with only one measurement distribution.
For example if you have a qubit in 50% |1>, 50% |0> and pass it through CNOT with a qubit in |1>. You get:
|11> 0
|10> 50%
|01> 50%
|00> 0
But if the second qubit was in state |0>, you get |11> 50%
|10> 0
|01> 0
|00> 50%
> if I have several pairs of entangled electrons all with spin of 50% up / 50% downAlso just in case you don't know qubits have phase so there is more than one way to have a qubit that when measured will be up 50% of the time.
The state of the system could be, for example:
State Amplitude Probability
+/+ 1/2 1/4
+/- -1/2 1/4
-/+ -i/2 1/4
-/- i/2 1/4
In the right-hand column, I've included the probability of measuring each state (notice that it's the square of the modulus of the amplitude), but that's not the fundamental quantity that you deal with in quantum mechanics.All the dynamics of a quantum system are expressed in terms of how the amplitudes change over time. In fact, if this means anything to you, the most succinct way to state how quantum mechanics works is that a quantum system with N possible states is represented by an N-dimensional complex vector, and that in an infinitesimal time step dt, the system goes from v to (1+iHdt)v, where H is a matrix with complex entries (and 1 stands for the identity matrix). If you measure the quantum system, you have to first pick a set of basis vectors in which to measure it. The result you get is one of the basis vectors. The probability of getting any basis vector as result is the square of the coefficient (technically, the square modulus of the coefficient) on that basis vector. The coefficient is what we call the "amplitude."
In a quantum computer, you first prepare the system in a desired state (a complex vector). Then, you get to choose what linear operations you will apply to the state. Then, you observe the state in some basis (in the complex vector space), and get a random basis vector (proportional to the modulus squared of the coefficients in the final state). That's basically it. The trick is whether or not you can actually figure out a way to compute anything useful with such a system with lower complexity than you can with a classical computer. Your fundamental operations are different (linear transformations of a complex vector) than they are in a classical computer, and you have the added wrinkle that you can't read off the final state of the computer - the final result is a random draw from a probability distribution that is based on the state of the computer. Peter Shor figured out that with this setup, you can factor large numbers in a way that uses fewer operations than a classical computer (asymptotically). It also turns out that you can simulate quantum systems really effectively with this sort of system. That's not so surprising. As Feynman said,
> "Nature isn't classical, dammit, and if you want to make a simulation of Nature, you'd better make it quantum mechanical, and by golly it's a wonderful problem, because it doesn't look so easy."
The fuzzy results, in my understanding, are a property of quantum mechanics and don't require multiple qubits. A single electron has some probability of spinning up and a complementary probability of spinning down, and until you go ahead and measure the spin, the actual spin value hasn't coalesced to either of those states, instead being some fuzzy intermediate probability spectrum.
Entangling multiple particles means (again, as far as I understand) that their measurements cease to be independent -- if I have several pairs of entangled electrons all with spin of 50% up / 50% down, then I might expect the results of measuring the spin of those pairs to look something like this table:
when in fact, because these are entangled pairs, I will either get or What am I missing here?