Misconceptions/Hardware/The real cost of noise
BeginnerHardware6 minInteractiveVisual demo

The myth

Noise only slightly affects quantum results

01

Why people believe this

A 1% error rate per gate sounds small. Classical computers tolerate noise too — error correcting codes in memory, checksums in storage. Surely a little quantum noise just slightly degrades results.

02

The correction

Quantum noise is catastrophic in a way classical noise is not. A 1% depolarizing error per gate on a 100-gate circuit gives an overall fidelity of roughly 0.99^100 = 36%. The result is nearly useless. Classical bits can be copied and checked — quantum states cannot be cloned, so error correction requires enormous overhead. On NISQ devices with no error correction, even shallow circuits of a few dozen gates produce unreliable results.

03

Visual demonstration

Fidelity decay by error rate per gate

At just 1% error per gate, a 100-gate circuit retains only 37% fidelity. At 5% it collapses to near zero.

04

Try it in the simulator

What to do

Load the Bell state. Simulate with noise at 0% — perfect 50/50. Now drag the noise slider to 5%, then 10%, then 20%. Watch the distribution degrade toward a uniform mixture. At 20% error per gate the Bell state is nearly destroyed after just 3 gates.

Open in simulator
05

Research notes

Tags

#noise#decoherence#error correction#NISQ#fidelity

Related cases

← Back to the misconceptions section