This is a small exercise, but it is a good one. Take a general spin-1/2 state ∣Ψ⟩=(αβ) (with ∣α∣2+∣β∣2=1). What is the probability of measuring the spin positive along y, and along z? To answer it we need the three things one uses again and again with spin-1/2: the Pauli matrices, rewriting a state in a different eigenbasis, and the Born rule. At the end we check the formulas on a few states where we already know what the answer has to be.
The Pauli matrix σy=(0i−i0) has eigenvalues ±1, which correspond to spin ±ℏ/2. Its normalized eigenstates are
∣↑y⟩=21(1i),∣↓y⟩=21(1−i)
Our state ∣Ψ⟩=(αβ) is written in the Sz eigenbasis, i.e. ∣Ψ⟩=α∣↑z⟩+β∣↓z⟩. To get probabilities along y we project onto the σy eigenstates instead.
The Born rule says P(sy=+2ℏ)=∣⟨↑y∣Ψ⟩∣2, which is the squared overlap with the eigenstate and nothing else (no operator sits in between). For the bra we need the conjugate transpose, ⟨↑y∣=21(1−i), and then
P(sy=+)=21(1−i)(αβ)2=21∣α−iβ∣2
A common slip
People sometimes write ∣⟨↑y∣Sy∣Ψ⟩∣2. Here that gives 4ℏ2 times the right answer, or exactly the right answer if you use σy. The second case only works because ⟨↑y∣σy=+⟨↑y∣ and ⟨↓y∣σy=−⟨↓y∣. It is a coincidence of eigenstates, not the Born rule.
The two probabilities should add up to one, and they do. The cross terms in the two moduli come with opposite signs and cancel, so P(+)+P(−)=21(∣α−iβ∣2+∣α+iβ∣2)=21(2∣α∣2+2∣β∣2)=1 ✓
Nothing surprising here. We are using the Born rule in the basis the state was already written in, so α and β are directly the probability amplitudes for sz=±ℏ/2.
A state that is spin-up along z is equally likely to be found up or down along y. This is what one expects, since Sy and Sz do not commute ([Sy,Sz]=iℏSx=0), and an eigenstate of Sz has the maximum possible uncertainty in Sy.