What does standard randomized benchmarking estimate?

Question: Under the usual assumptions of standard randomized benchmarking, what is extracted from the decay of survival probability with sequence length? A) The diamond-norm distance of every individual gate from its ideal implementation B) A complete process matrix for the average noise channel, including all SPAM parameters C) The worst-case coherent error accumulated by the longest benchmark sequence D) An average error parameter for the sampled gate set, with state-preparation and measurement effects largely absorbed into fit constants Correct: D Explanation: Randomized benchmarking fits an average decay curve whose decay parameter is related to an average gate-set error rate under suitable assumptions. SPAM errors mainly affect the offset and scale of the decay curve. The result is not complete process tomography and does not directly give a worst-case diamond-norm error. Topic: advanced quantum computing / randomized benchmarking / characterization