The electrical restitution curve (ERC) traditionally describes the recovery of action potential duration (APD) as a function of the interbeat interval or, more correctly, the diastolic interval (DI). Often overlooked in modeling studies, the normal ventricular ERC is triphasic, starting with a steep initial recovery at the shortest DIs, a transient decline, and a final asymptotic rise to a plateau phase reached at long DIs. Recent studies have proposed that it would be advantageous to lower the slope of the ERC by drug intervention, as this might reduce the potential for electrical alternans and ventricular fibrillation. This review discusses the pros and cons of a flat versus steep slope of the ERC and draws attention to mechanisms thatjustify the (physiologically) steep slope, rather than a flat slope, as a better design against arrhythmias. Five potential mechanisms are discussed, which allows for a different interpretation of the effect of the slope on arrhythmogenicity. The most important appears to be the physiologic rate adaptive shortening of APD that, by reciprocal lengthening of the DI, allows the subsequent APD to move more quickly from the steep initial ERC phase onto the flat phase. A less steep initial ERC phase would protract the transition toward more fully recovered APD and, in fact, may perpetuate electrical alternans. The triphasic ERC time course in normal myocardium cannot be explained by or fitted to single exponentials or single ion channel recovery kinetics. A simple tri-ionic model is suggested that may help explain the shape of the ERC at various repolarization levels and place APD recovery into perspective with intracellular calcium recycling and recovery of contractile force.