Dynamic Alignment as a Statistical Survival Effect

Dynamic alignment in magnetohydrodynamic (MHD) turbulence is usually taken to mean that Els\"asser fluctuations become increasingly aligned at smaller inertial-range scales. We show that this is not the correct physical interpretation of standard weighted measurements: the measured small angles result from statistical selective survival rather than from a volume-filling tendency toward alignment. Our framework separates angular statistics from Els\"asser-amplitude weighting and describes amplitude--angle populations through selective retention. Large-amplitude large-angle events deplete faster than large-amplitude small-angle events, so the latter are overrepresented in weighted measurements although the typical folded angle shows no progressive scale dependence. The aligned strong-event population need not be produced more efficiently; it can dominate second-order amplitudes because it has a longer residence time. Thus retention allows effective Els\"asser-increment scaling shallower than $\ell_\perp^{1/3}$, without scale-dependent alignment of the typical angle. We test this using JHTDB MHD data and Wind observations. In JHTDB, the unweighted folded angle remains moderately below the random baseline, while weighted diagnostics sample intense Els\"asser-increment events whose amplitudes correlate with smaller folded angles. The measured retention balance produces effective Els\"asser-increment scaling close to $\ell_\perp^{1/4}$, giving $k_\perp^{-3/2}$ rather than Kolmogorov--Richardson $k_\perp^{-5/3}$, while the typical folded angle shows no progressive scale-dependent alignment. Wind shows the same angle--amplitude hierarchy and negative covariance. Dynamic alignment, as usually measured, is thus better understood as selective survival of small-angle intense fluctuations, with a nearly scale-independent typical angle and a $k_\perp^{-3/2}$-type effective spectrum.