Gait timing dynamics of treadmill and overground running were compared. Nine trained runners ran treadmill and track trials at 80, 100, and 120% of preferred pace for 8 min. each. Stride time series were generated for each trial. To each series, detrended fluctuation analysis (DFA), power spectral density (PSD), and multiscale entropy (MSE) analysis were applied to infer the regime of control along the randomness-regularity axis. Compared to overground running, treadmill running exhibited a higher DFA and PSD scaling exponent, as well as lower entropy at non-preferred speeds. This indicates a more ordered control for treadmill running, especially at non-preferred speeds. The results suggest that the treadmill itself brings about greater constraints and requires increased voluntary control. Thus, the quantification of treadmill running gait dynamics does not necessarily reflect movement in overground settings.