Create a small measurement notebook around a two-qubit Bell circuit. Start in a simulator or circuit composer. Save predictions, counts, and comparisons at each stage.
Use the same 5-minute recall, 20-minute practice, 5-minute reflection loop. Move forward when the deliverable and failure checks work. Real hardware is optional and may have access or usage limits.
Six stages, one experiment notebook
01 / CLASSICAL BASELINEPredict two independent bits
Enumerate 00, 01, 10, and 11 for two fair, independent coins. Generate or tally 100 samples and compare counts with the expected 25% each.
Deliverable: a count table and a sentence explaining sampling variation.
Failure checks: counts sum to the number of trials; no negative counts; changing the random seed changes counts without changing the expected probabilities.
02 / ONE QUBITPrepare and measure
In a simulator, measure a qubit prepared as 0. Add an X gate and repeat, then replace X with H. Predict each outcome distribution before running it.
Deliverable: three circuit sketches and 100-shot count tables.
Failure checks: ideal 0 and X preparations are deterministic; H outcomes approach 50/50 over many shots; one measurement does not reveal both amplitudes.
03 / BELL CIRCUITCreate correlated outcomes
Apply H to the first qubit, then a controlled-X from the first to the second. Measure both in the computational basis. Compare with your independent-coin baseline.
q0: H → control → measure
q1: X → measure
Deliverable: a circuit screenshot, predicted probabilities, and counts for 1,000 shots.
Failure checks: verify the control and target; document bit order; the ideal circuit yields only 00 and 11 in this basis.
04 / STRONGER CLASSICAL COMPARISONChallenge your explanation
Use one fair classical coin to output either 00 or 11. This shared-randomness baseline can match the Bell circuit’s computational-basis counts. Explain why those counts alone do not establish entanglement.
Read about measurements in other bases before extending the experiment. A proper entanglement witness or Bell test needs a defined protocol and assumptions.
Deliverable: a side-by-side table for independent coins, shared randomness, and the circuit.
Failure checks: do not claim a speedup or prove entanglement from one measurement basis; check each method uses the same number of trials.
05 / NOISE AND REPEATED RUNSMeasure variation and error
Repeat the ideal experiment at 100 and 1,000 shots. If your simulator supports a noise model, add a documented small error probability. Hardware is an optional comparison.
Deliverable: repeated count tables, the fraction of 01/10 outcomes, and the noise settings.
Failure checks: distinguish sampling variation from simulated noise; label ideal, noisy, and hardware results; record the device and run conditions.
06 / EXPLAIN AND SHAREPublish a reproducible note
Collect your circuit, baseline code or tally method, predictions, results, and limitations. Include one surprising result and a test you would run next.
Deliverable: a public notebook or project note another learner can follow.
Failure checks: remove account tokens; verify that counts and plots agree; state clearly that this experiment demonstrates no computational advantage.
Use the showcase submission checklist ↗