intermediate · Physics · The Postulates of Quantum Mechanics
Expectation Values, Formally
An expectation value is the average result you would get by measuring an observable on many
identically prepared copies of a state. It is not (in general) one of the possible outcomes — it is
their probability-weighted mean. This lesson derives the formula from the Born rule and computes one
by hand.
From outcomes to averages
Measuring A^=∑kakP^k on ∣ψ⟩ gives outcome ak with probability
p(ak)=⟨ψ∣P^k∣ψ⟩. The average outcome is therefore the ordinary
statistical mean,
⟨A^⟩=k∑akp(ak)=k∑ak⟨ψ∣P^k∣ψ⟩.
Now use the spectral decomposition A^=∑kakP^k to pull the sum inside the
expectation:
⟨A^⟩=⟨ψ∣(k∑akP^k)∣ψ⟩=⟨ψ∣A^∣ψ⟩.
This is the compact and famous form of the expectation value:
⟨A^⟩=⟨ψ∣A^∣ψ⟩
The two expressions are identical — the projector form makes the statistical meaning explicit, while
⟨ψ∣A^∣ψ⟩ is the convenient form for calculation.
A worked qubit example
Take ∣ψ⟩=cosθ∣0⟩+sinθ∣1⟩ and the observable Z^,
which has eigenvalue +1 on ∣0⟩ and −1 on ∣1⟩. The Born-rule probabilities are
p(0)=∣⟨0∣ψ⟩∣2=cos2θ,p(1)=∣⟨1∣ψ⟩∣2=sin2θ.
So the expectation value is
⟨Z^⟩=(+1)cos2θ+(−1)sin2θ=cos2θ−sin2θ=cos2θ.
For θ=π/6 this gives cos(π/3)=21. You can confirm with the matrix form: