What does an RB decay parameter not determine?

Question: Two gate implementations have the same average infidelity inferred from randomized benchmarking. Which conclusion is generally unjustified? A) Their fitted RB decay rates may be similar under the benchmark model B) Their worst-case errors and long-circuit behaviour must therefore be identical C) State-preparation and measurement errors may primarily alter fit constants rather than the principal decay rate D) One implementation could have more coherent noise than the other despite a similar average error Correct: B Explanation: Average gate fidelity does not uniquely specify the structure of the noise. Two channels with similar average error may have very different coherent components, diamond-norm distances, temporal correlations, and cumulative behaviour in algorithmic circuits. Topic: advanced quantum computing / benchmarking / coherent errors