Cranky loveable geezer, yes, but ungracious sourpuss, I wouldn't have thought! Be that as it may, I think Dennis' basic point stands since the substantive part of your response to him (i.e., the stuff that does not get into micro-analysing the data) is unconvincing because of its use of "intention" as a meaningful criteria. The distinction between "pure" accident (and we already know in many cases these are not "pure" accidents but cases of "cost-cutting") and homicide can be made relevant only if:
(a) the workers who face "pure" accidents are unaware of this data, in which case a disingenuous claim can be made that they are not therefore "really" laying their lives on the line.
(b) if we subscribe to your idea that homicides (intentional acts) are a more serious risk than (in your words) unavoidable pure accidents. But the opposite is probably true. The intentional acts, I would argue, are easier to predict, plan for and avoid, especially in an environment that valorises the person taking them on, than pure and unavoidable accidents.
--ravi