We studied the torsion behavior of α-Fe nanowires seeded with twin boundaries (TBs) using molecular dynamics simulations. Twisting the wire generates topological defects in the twin walls, namely kinks inside the twin walls for small twist angles, and junctions between kinks for large twist angles. During twisting the kink motion is jerky and uncorrelated at small twist angles. The probability density function (PDF) of jerk energies follows approximately a Gaussian distribution, indicating a mild deformation mode. The kink dynamics transforms from mild to wild at larger twist angles when complex twin patterns with a high density of junctions are generated. The collective motion of kinks now shows avalanche behavior with the energy being power-law distributed. The wildness, which measures the proportion of strain energy relaxed through such avalanches, is correlated with the junction density, and controlled by the external length scale (wire diameter) as well as an internal length scale (twin boundary spacing). Good strain-stress recoverability is achieved when unloading the wire before the formation of complex twin patterns. We correlate the evolution of twin patterns with a statistical analysis of jerk dynamics, which identifies the unique mechanical properties governed by twin boundary motion in nanowires.