every sampled payload reproduced the same directional pattern. With no clipping, low-correlation heavy tails generated very large false spikes, for example false_spike_excess about 136.5 at nu=3, rho=0 and about 70.2 at nu=3, rho=0.6; clipping drove those excesses sharply down, to about 7.60, 3.93, and 1.85 at rho=0 for k=3.5, 2.5, and 1.8 respectively, while the strongest-correlation cases showed mixed alignment changes and increasingly negative contrast entries for aggressive clipping.