I just dont think ur math adds up. Here’s my logic on the matter

Loss of liberty = population * smoker% * average liberty of someone to smoke a ciggy

Gain in liberty = population * (1- smoker%) * relative air quality improvement * average liberty of someone to not be exposed to ciggy smoke

I would simply argue the relative air quality improvement is so small that the gain in liberty will never be greater than the loss. Lets assume that ciggy smoke distributes according to inverse square law and smoking reduces ur lifespan by 50% at a distance of 1m ur life is reduced by 0.039% at 2m its reduced by 0.0099% for comparison by my rough estimation the avergae driver loses about 0.036% of their life by driving and society accepts that risk so all u need to do is stand a little over a metre away from anyone with a ciggy and its safer than driving (i did all the calcs in favour of ciggies being as bad as possible.

This is all just fermi estimation but we should be in the correct order of magnitude.