| > If you say half of drivers are not at fault, it's only because we assume only one person can be at fault. You're absolutely right; there are at least some collisions that are the fault of both/multiple drivers, or of no drivers. I did assume that fault is most often one driver in two-car collisions (and I still think that), but I shouldn't have stated it as fact, I was thinking of it as the premise of the subsequent point. My apologies. > Most collisions involve more than one mistake That's a strong claim you're making. Citation needed. Here are some resources: https://crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/... http://www.iihs.org/iihs/topics/t/general-statistics/fatalit... > it's certainly possible for a sufficiently good driver to improve their odds by much better than half Of course. Single vehicle crashes are already more than half. 30% of all car fatalities have been attributed to drunk driving. Don't ever get in a car after drinking, and your odds are already much better. Another 30% of car deaths occur when speeding. Stay under the limit, and your odds are that much better again. I hope you can see that I wasn't claiming that half is the actual limit, I was making a point about the arithmetic of probabilities. My point is that if factors out of your control are responsible for x% of crashes, then the maximum upper limit of good you can do by honing your skill and being the best driver in the world is 100-x%. The point is there's an upper bound to what training and preparedness can do for you, and that upper bound can always be exceeded by avoiding the risk in the first place. |