[vundercind]

Sure, I agree that 800,000-hospital-visits number isn’t really a great picture of what’s going on. The deaths number was just so very far under what I was sure was the serious-harm figure that I thought it worth bringing in the non-death attacks, and a (probably reasonable-ish) estimate of cases that prompted treatment was first thing I saw that looked close to what I was looking for.

The 17.5k stat’s interesting—I think you’re right that it may at least be in the ballpark for an estimate of unrestrained dog attacks. Some would generate a claim, some wouldn’t, some claims wouldn’t be for unrestrained dogs… yeah, probably a good starting point. I like that one, good eye.

I’d guess most attacks of that 800k aren’t from strangers’ dogs at all, but friends and family’s dogs. Simple matter of opportunity and time exposed.