Risk is the chance of something happening coupled with the effect it would have. Perhaps the police should not be the ones protecting the public from a Mumbai style attack - possibly it should be the military - but it should be someone because of the grave consequences.
That does not have any link to whether police should routinely use de-escalation techniques (which they should).
There has never been a Mumbai-style terror attack, and the National Guard could be activated for such a thing.
Some edge cases deserve more addressing than others.
Not only is no one arguing different, the "hundreds of thousands of car accidents in the USA annually" has been addressed (in several ways). Perhaps other problems may also be addressed?
> There has never been a Mumbai-style terror attack
No, only in most (if not all) of the other G8 countries, but have they had a Pulse nightclub attack or planes flown into buildings? How many have had a Bataclan or a London Bridge style attack?
Let's not split hairs.
> the National Guard could be activated for such a thing
Well, I did suggest “possibly it should be the military” so the National Guard, being part of the military, would fit.
In short, an edge case being rare does not imply that it should not be prepared for. What's the response time for the National Guard if terrorists start attacking people in the street?
Edit: typo
'Hours' would probably be quick, since you'd probably need to get the governor's approval first.
a) whether they are edge cases, which they are in terms of occurrence b) risk, which are comparable on a personal level
Additionally, when seat belt laws were introduced there were arguments about whether it was a moral hazard. Their requiring some effort (try strapping in a small child) and considerable danger to some (again, children, and pregnant women) was not relevant to that (ultimately wrong) argument and it's not relevant here either.