Obesity remains a major public health concern and intermittent fasting is a popular strategy for weight loss, which may present independent health benefits. However, the number of diet books advising how fasting can be incorporated into our daily lives is several orders of magnitude greater than the number of trials examining whether fasting should be encouraged at all. This review will consider the state of current understanding regarding various forms of intermittent fasting (e.g. 5:2, time-restricted feeding and alternate-day fasting). The efficacy of these temporally defined approaches appears broadly equivalent to that of standard daily energy restriction, although many of these models of intermittent fasting do not involve fed-fasted cycles every other 24 h sleep–wake cycle and/or permit some limited energy intake outside of prescribed feeding times. Accordingly, the intervention period therefore may not regularly alternate, may not span all or even most of any given day, and may not even involve absolute fasting. This is important because potentially advantageous physiological mechanisms may only be initiated if a post-absorptive state is sustained by uninterrupted fasting for a more prolonged duration than applied in many trials. Indeed, promising effects on fat mass and insulin sensitivity have been reported when fasting duration is routinely extended beyond sixteen consecutive hours. Further progress will require such models to be tested with appropriate controls to isolate whether any possible health effects of intermittent fasting are primarily attributable to regularly protracted post-absorptive periods, or simply to the net negative energy balance indirectly elicited by any form of dietary restriction.