Table 1

Linear regression on instantaneous head position (H) as function of the head-position error (ΔE; Eq. 11b), and on head velocity (H˙ ) as function of the velocity error (ΔE˙ ; Eq. 11a)

Regression	Offset	Slope	Correlation
H=α+β·ΔE	−0.98	−1.13	0.55
H˙=ρ+η·ΔE˙	0.00	−0.96	0.63
H=a+b·ΔE˙	0.16	−0.009	0.01
H˙=c+d·ΔE	0.20	0.15	0.03

Both regressions yield high gains and correlations. Head position is unrelated to the velocity error, and, conversely, head velocity is independent of the position error. All regressions were performed on a subset of 48,400 data points (trajectories were sampled in steps of 15 samples, because of memory limitations).