In part for the apparent increased electrical consumption and noise but mostly because of the following:
-relatively confined waters
-variable winds causing radical changes in apparent wind direction
-variable and frequently changing current effects
-motion of the masthead fly tends to affect the apparent wind readout and thus the autopilot response disproportionately to actual heading adjustments req'd (despite damping)

If I were offshore, I'd be much more motivated to find a good solution.