Alice holds an unknown qubit C = a|0⟩ + b|1⟩ (set by θ) that she wants to send to Bob, plus half (A) of an entangled Bell pair she shares with Bob's half (B). Alice performs a joint Bell-state measurement on C and A. That measurement can only ever return one of four equally likely outcomes — it never reveals a or b directly, so C's state is destroyed, not copied.
|Φ+⟩ = (|00⟩+|11⟩)/√2 (shared pair)
outcome ∈ {00, 01, 10, 11}, each p = 1/4
Bob's raw qubit = a gate applied to (a,b)
Alice sends her 2-bit outcome to Bob over an ordinary classical channel — this is the step that takes real time (the delay slider) and is why teleportation can never send information faster than light. Once Bob has the bits, he applies exactly one of four corrections to his half of the pair:
00 → I (do nothing)
01 → X (bit flip: a,b → b,a)
10 → Z (phase flip: a,b → a,-b)
11 → XZ (both: a,b → b,-a)
- θ — the state being teleported, drawn as an arrow on the amplitude circle; watch it reappear on Bob's qubit only after the correction lands.
- Channel delay — how long the 2 classical bits take to cross from Alice to Bob; teleportation is instant only in the trivial sense that entanglement correlates, never in the sense that usable information moves faster than this delay.
- Channel noise — models a decohered Bell pair, the real-world limitation the article calls out: fidelity drops below 100% even though the protocol steps are unchanged.
This is exactly the protocol described in "Quantum Teleportation Explained": entanglement sets up the correlation, the Bell measurement destroys the original while extracting two classical bits, and Bob's conditional gate is the only step that actually reconstructs the state.