Setting aside whether reverse risk compensation actually works (and it would be absolutely bizarre to have >100% risk compensation), Tullock's spike has the problem that it makes driving less effective, i.e. people get from point A to point B more slowly. If your policy goal is for people to drive less then sure. If your goal is for people to be safer in the course of achieving their actual objective of getting from point A to point B, though, the policy is spectacularly bad.
That's really not the point. The thought experiment doesn't consider the safety, and far less the convenience (did you call this "effectiveness"?), of drivers. Rather, it examines the safety of all those who suffer the presence of drivers: pedestrians, cyclists, road workers, wild animals, unleashed dogs, etc.
Sure, that is a goal for many drivers. There's no reason for anyone else to care about it. Whatever the hypothetical advantages of automobiles to society in general, police-enforced speed limits exist. Since you seem opposed to limits on speed, one might expect you to value Tullock's spike as another possible solution in the trade-off space that would allow faster travel for those who, like you, really value that, while still somewhat protecting the rest of us on the street.
