One of the most effective diets I'm aware of is the coffee and cigarette diet. Anytime you feel hungry, have a coffee and a cigarette first, then eat a little if you're still hungry.
It has a much better compliance rate that other calorie-restriction diets.
Smoking increases hazard rates by 3x. Even the extreme obesity (obesity III, formerly morbid obesity) increases hazard rates less than this. So the risk tradeoff for picking up smoking in favor of reducing obesity is not good. It's like saying you'll stop wearing a seatbelt to reduce the risk of being trapped in a burning car; your overall risk profile is still higher.
Or is it that you don't think health is the goal of reducing obesity? If so, most of the metrics we use to deem obesity as sub-optimal are wrong, including those in TFA. Better compliance to a diet that increases overall health risk is not a good strategy IMO.
It has a much better compliance rate that other calorie-restriction diets.