Interventions and policies designed to reduce tobacco use differ according to the population they are intended to reach and the nature of the behavior they are designed to alter. Nicotine replacement therapy is primarily intended to help older smokers to quit. Laws restricting tobacco sales to minors prevent youth from starting to smoke. Anti-tobacco media campaigns can be targeted at people of any age or gender and can convey any message: don't start, quit now, or don't relapse. Because these interventions affect different behaviors in different populations, their ultimate impact on the nation's health will vary. OBJECTIVE: We estimate and compare the relative magnitude of the public health gains from preventing smoking initiation vs. encouraging cessation or avoiding relapse in different ages and genders. METHODS: We define health gains as the predicted increase in quality-adjusted life-years (QALYs) to the US population over a century. To estimate QALYs, we developed the Tobacco Policy Model. The model simulates a 10% reduction in the annual probability of initiation vs. a 10% increase in cessation vs. a 10% reduction in relapse in males and females in six age groups: 10-19, 20-29, 30-39, 40-49, 50-59 and 60-69. RESULTS: Among youth and young adults, reducing initiation yields far more QALYs than encouraging cessation or averting relapse. In middle-aged adults, cessation yields the most QALYs, followed by averting relapse and reducing initiation. In the oldest age group, averting relapse yields the most QALYs followed by cessation prevention and reducing initiation. In general, increasing cessation and reducing relapse is more beneficial in males than in females while reducing initiation is more beneficial in females. CONCLUSION: The relative value of preventing initiation, encouraging cessation and averting relapse differs by age and gender. Reducing initiation in youth is likely to offer the largest public health impact over the next century.