One of my favorite classes in college was statistics. One of the problems involved a traffic study. I recreated it as an example.
I'm living off of one of the most notorious intersections in Denver CO right now, Parker rd. and Quincy Ave. Anecdotally, I hear sirens 3-4 times a day. At 2-3 cars on average per accident, that's call it on average 10 cars a day involved in an accident PER DAY, some involving injuries, and yes, some even fatal. Best avoid that death trap, wouldn't you say?
And yet, I just looked up the numbers, and on average 100,000 cars pass through that intersection per day. So 99,990 passed through without incident, 99.99% of them. Now, if my brother in law had been in 2 accidents there, anecdotally and emotionally I would definitely be less inclined to pass through that intersection... but based on statistics it would be ridiculous to close the intersection based on the numbers.
From a statistics standpoint, we have to include risk factors. Of the 10 cars involved in the accidents, we can make a reasonable extrapolation that 5 of them were at fault in some way. Those we can call the at risk drivers. Now, looking at the volume of traffic, we can suppose that, being generous, there were another 90 close calls that didn't result in a crash, for a total of 100 incidents per day, just to make the math easy. So 99,900 cars made it through the intersection without incident, or 99.9 percent.
If we carry out this analogy to trees, and honestly after cleaning up several hurricanes and countless storms and wind events I think the numbers aren't far off, 99% or so of trees will be just fine, and the 1% are generally high risk trees, whether through age, disease, isolation, or some other mitigating risk factor. The proximity to a structure shouldn't be the only factor considered. Anecdotally it sucks when it happens, but statistically it shouldn't be the only factor during risk assessment. Using that logic, statistically Houston should be a prairie, and yet, probably one of the most heavily wooded metro areas in the nation.