Same here. Ian's mistaken if he thinks primary is the solution to everything. He's right that it enhances the odds considerably, but it doesn't solve all problems.
Primary isn't the solution to everything (I don't think I ever said it was). But proper lane positioning (which includes the use of primary when appropriate) combined with careful route selection (and I don't mean avoiding busy roads), proper speed for the road/intersection visibility situation ahead and adequate lighting/visibility for low light situations is the solution to everything that can be solved by the cyclist. I stress primary, because too many people are scared to use it, when it can, as you say, greatly reduce problems.
Having said that, there is always an element that cannot possibly be addressed by the cyclist, but getting more than one close call per month suggests to me that the cyclist isn't doing all he can to prevent problems. Like I said, I haven't had a serious problem in over a decade, and I certainly don't get close calls every week or even every month or every few months, so I must be doing something right - it can't be just luck. Anyone measuring incidents per mile is, in my view, doing something very wrong - it can't just be lots of bad drivers, because the careful cyclist who is in control of his lane can control bad drivers and mitigate how much their incompetence affects him.
Parking lots, however, are a different issue. Visibility can be virtually nil in parking lots, and people just blindly pull out or speed up to get a spot, so you just have to be really careful. Many problems, I think, come from cyclists insisting on riding too fast for the visibility/potential-threat situation, and this applies especially to parking lots.