Casino-style self-exclusion presents a physical barrier to further betting, and also lowers the reward to gambling, through the threat of confiscation of winnings (and embarrassment at being caught in violation of the exclusion order). Now-standard(?) commitment contracts implement the reward diminution element, but not the physical unavailability, when it comes to refraining from the undesired behavior.
A recent article by Gine', Karlan, and Zinman (working paper version here) examined a field trial in the Philippines of commitment contracts aimed at smoking cessation. Smokers were offered a chance to conveniently make weekly deposits to an illiquid bank account. (The idea was that they would deposit about the amount they would otherwise spend on cigarettes.) After six months, the depositors who took up this option would be tested to see if they had quit smoking. If they had quit, they would have access to their funds; if not, they would lose the money. The vast majority of smokers who were offered the savings account option refused to take part. Most of those who did take up the commitment contract failed to quit smoking. Nevertheless, quit rates were higher for those offered the savings account option than for those in the control group, and the difference persisted six months later.