Human annotators penalize an agent's extra conversation turns twice as hard as extra tool calls.

RideWay pairs a stateful tool-calling ridehailing benchmark with Efficiency Utility, a success-gated metric discounting trajectories for excess tool calls and user-facing turns against task-specific reference effort, penalties calibrated from human paired preferences. Across 58 tasks and 24 models the fitted penalty for excess turns is about 2x that for excess tool calls. On held-out preferences the metric hits 78.7% accuracy overall, 90.6% when trajectories differ in turns, and chance level when they differ only in tool calls. Asking the user one extra question costs you more than two extra searches.