Speed Over Strategy: Why Agent Velocity Dominates Aggregation Method in Crowd-Sourced Continuous Control

Crowd-sourced control systems—in which distributed participants collectively operate a shared agent through real-time voting—are formalized in robotics as Multi-Operator Single-Robot (MOSR) teleoperation and instantiated in audience participation platforms. The central design question is how to aggregate participant inputs into continuous agent movement; yet no prior work has systematically compared the relative importance of aggregation method versus agent operating speed. Through 8,010 simulation runs, a two-way ANOVA reveals that agent speed accounts for 46–97% of tracking error variance while aggregation method accounts for only 2–29%—a ratio of 1.6 to 44× (median 6.9×). Within tested smooth continuous trajectory regimes, speed dominance is consistent across adversarial participant ratios of 5–50%, and post-hoc analysis indicates that methods with identical effective speeds show no significant performance difference. A trajectory-dependent optimal speed v* ≈ 2.0 m/s (2.03 ± 0.03 m/s across tested conditions) remains invariant to adversarial ratio. A crowd-size ceiling effect is documented where improvement plateaus beyond a critical group size. These findings are unified in a four-parameter multiplicative model, RMSE(v, N, p) = f(v) × g(p, N), achieving R² = 0.984 and 4.6% out-of-sample error. The key design guideline: optimize agent speed first, then scale participation to the point of diminishing returns.