I agree with you that the speed should be much reduced in inclement weather. Other significant hazards can't be predicted as easily: drivers veering into the op

I agree with you that the speed should be much reduced in inclement weather. Other significant hazards can't be predicted as easily: drivers veering into the opposite lane; moose; debris on the road; people running red lights, etc. In all of these cases, driving a bit slower can result in significant reductions for both occurrence rate (extra reaction time) and severity (kinetic energy being proportional to v squared). The statistic that most accidents occur under the speed limit doesn't say much. (I'm not aware of the study you refer to, but I suspect a good deal of the phenomenon is due to intersection collisions.) What you want is the converse - driving under the speed limit causes most accidents (I doubt it). And even that would not be sufficient: what counts is the number of fatalities, and there are probably fewer deaths in low-speed accidents.