Anticoagulation for stroke prevention in atrial fibrillation (AF) is effective. Pivotal trials RE-LY, ROCKET AF, ARISTOTLE, and ENGAGE-AF TIMI 48 tested novel agents against warfarin (W). In RE-LY, an open-label trial, dabigatran 150 mg BID (D150) was superior (35%) and 110 mg BID (D110) was noninferior to W. D150 reduced ischemic strokes by 25% and intracerebral bleeds by 74%, but increased major GI bleeds by 0.5 % per year. In ROCKET AF, a double-blind study, rivaroxaban 20 mg daily, downtitrated to 15 mg daily (if CrCl was <49) was noninferior for efficacy and safety, with an increase in GI bleeds. In ARISTOTLE, a double-blind study, apixaban 5 mg BID (downtitrated to 2.5 mg BID if two of the following were present: age, >80; weight, <60 kg; or serum creatinine, >1.5 mg) was superior for safety (31%), efficacy (21%), and all-cause mortality (11%). In ENGAGE-AF TIMI 48, edoxaban 60 mg once daily (30 mg once daily if CrCl 30-50 ml/min, weight <60 kg, or concomitant verapamil or quinidine) was noninferior to W for efficacy, but reduced major bleeding (20%). To translate clinical trials to practice, understanding the disease and each anticoagulant is essential. For all novel agents, rapid anticoagulation, absence of monitoring, and a short half-life differentiate them from W. Bleed rates were either noninferior or lower than for W, without an antidote. Patient compliance is critical. Knowledge of renal function is essential and maintaining patients on therapy is key.