Vectors and superposition as weighted sums

The word “superposition” sounds mystical. It’s not. It’s the same thing you did in algebra class: a weighted sum.

This lesson is an attitude adjustment. The goal is to make the word “superposition” feel as boring as the word “combination” — because that’s what it is.

Vectors you already know

Think about a 2D plane. You can describe any point on it with two numbers — $x$ and $y$ coordinates. But you can also describe it as a weighted combination of two special “basis” vectors:

$\vec{e_x} = (1, 0), \quad \vec{e_y} = (0, 1)$

Any point $(a, b)$ on the plane is just:

$a \cdot \vec{e_x} + b \cdot \vec{e_y}$

Nothing strange here. If I want the point $(3, 4)$, I say “three units of $\vec{e_x}$ plus four units of $\vec{e_y}$.” Done.

This is a superposition — specifically, a superposition of $\vec{e_x}$ and $\vec{e_y}$ with weights $3$ and $4$. We just don’t usually call it that in high school math.

Qubits are the same idea

A qubit state is literally the same structure. Instead of points on a plane, we have “abstract points” in a 2-dimensional complex space. Instead of basis vectors $\vec{e_x}$ and $\vec{e_y}$, we have basis kets $|0\rangle$ and $|1\rangle$. And instead of real weights $a$ and $b$, we have complex weights $\alpha$ and $\beta$:

$|\psi\rangle = \alpha \, |0\rangle + \beta \, |1\rangle$

That’s it. That’s the whole “mystery” of superposition: a qubit state is a weighted combination of the two basis states, where the weights are complex numbers.

The only extra constraint, compared to Euclidean vectors, is the normalization condition:

$|\alpha|^2 + |\beta|^2 = 1$

This makes sure the probabilities add up to 1 when we measure. No corresponding rule exists for arrows on a piece of paper — they can be any length they like. For qubits, the length is pinned to 1 by physics.

Three different pictures, one thing

You will see the same qubit state written in several different ways throughout this course. They are all saying the same thing. Get comfortable translating between them.

Dirac notation:

$|\psi\rangle = \alpha|0\rangle + \beta|1\rangle$

Column vector:

$|\psi\rangle = \begin{pmatrix} \alpha \\ \beta \end{pmatrix}$

Coordinate pair:

$|\psi\rangle \leftrightarrow (\alpha, \beta)$

They are the same object. The Dirac form is good for algebra. The column-vector form is good for matrix multiplication (which gates are). The coordinate pair is good when you’re just listing things in code.

Why “superposition” is not the same as “probabilistic”

A classical probabilistic bit that is $|0\rangle$ 50% of the time and $|1\rangle$ 50% of the time is not in a superposition. It’s in a mixture. The difference is subtle but important: in a mixture, the bit is either $|0\rangle$ or $|1\rangle$ — we just don’t know which. In a superposition, the qubit is genuinely $\alpha|0\rangle + \beta|1\rangle$, with both components coexisting until measurement.

The cleanest way to see the difference: mixtures can’t interfere, but superpositions can. Two paths leading to the same measurement outcome can have amplitudes that cancel (destructive interference) or reinforce (constructive interference) — something classical probability simply doesn’t do. Negative probabilities make no sense; negative amplitudes make total sense.

You’ll see this happen for the first time when we apply two Hadamard gates in a row and get back to where we started.

What’s next

You now have bits, probability, complex numbers, and vectors. In Module 2, we meet the qubit for real — and from that point forward, the picture takes over. Every concept gets a widget. The math will still be there when you need it.