Current life-history paradigms predict that an organism's total reproductive output is determined by its intrinsic level of ‘reproductive effort’ (RE), interacting with local resource availability. In turn, RE is driven by ‘costs of reproduction’. Within the total allocation, selective forces on offspring size ultimately determine both the number and size of offspring (because of a trade-off between these two traits). Although patterns of life-history variation are broadly consistent with this scenario, the evidence is weak: it is generally difficult to measure ‘costs’ or selective forces on offspring size. To test this paradigm, we need a study organism which differs in these respects from related taxa, so that we can unequivocally generate predictions about life-history traits. A remarkable island population of Asian pit-vipers Gloydius shedaoensis was studied for which (unlike mainland conspecifics) ‘costs of reproduction’ are low, food availability is high, and there is intense selection for large offspring size. Maternal body sizes and condition, and offspring numbers and sizes were quantified for 79 litters over 2 years. As predicted, the snakes display high reproductive output per female (relative to maternal body mass, litter mass is about twice as high as in related taxa), large offspring size (about three times the mass of related taxa) and a strong trade-off between offspring size and litter size. Reproductive output was only weakly correlated with maternal body size, and output was adjusted mostly by manipulating litter size rather than offspring size. Reproductive output varied between years, but maternal body condition after parturition did not. Measures of reproductive output (such as the mass or energy content of the litter) bear little relationship to ‘costs of reproduction’ in this system.