# Gate Fidelity
**Gate fidelity** measures how close a real quantum operation comes to the perfect operation it was meant to perform. A fidelity of $0.999$ means that, on average, the operation is right $99.9\%$ of the time and introduces an error $0.1\%$ of the time.
## First principles
A quantum **gate** is a controlled transformation of one or more [[Qubit|qubits]] — the quantum analogue of a logic operation. In practice every gate is imperfect: control pulses are slightly off, the qubit drifts, neighbours interfere. Fidelity $F$ quantifies the overlap between the state you actually produced and the ideal target state, so the **error per gate** is roughly $1 - F$.
> [!intuition] Errors compound
> If each gate has error $\varepsilon$ and an algorithm uses $N$ gates, the chance of getting through cleanly scales like $(1-\varepsilon)^N$. With $\varepsilon = 0.1\%$ you are in trouble after a few thousand gates. Real algorithms need *millions*, which is why raw fidelity alone is never enough.
## Why it matters
- **The threshold connection.** [[Quantum Error Correction]] only helps if gate fidelity is already above a critical value — the [[Code Distance and Threshold Theorem|fault-tolerance threshold]]. Below it, adding error correction makes things *worse*; above it, errors can be suppressed arbitrarily.
- **Two-qubit gates are the bottleneck.** Single-qubit gates are usually high-fidelity; the entangling two-qubit gates (e.g. via [[Rydberg Interactions]]) are harder and dominate the error budget.
- Fidelity must be weighed against **speed** and [[Coherence Time]]: a fast, slightly worse gate may beat a slow, perfect one if coherence is limited.
## Related
- [[Qubit]]
- [[Rydberg Interactions]]
- [[Coherence Time]]
- [[Code Distance and Threshold Theorem]]
- [[Quantum Error Correction]]