This talk discusses whether moral contractualists can solve the non-identity problem by appealing to the risk of a life not worth living. In a paradigmatic non-identity case, we can choose between creating a person with a good life and creating another person with a much worse but still positive life. It’s hard for contractualists to explain why we should, other things being equal, create the better-off person. Tim Henning (2022) has suggested a novel solution. He claims that if we deteriorate future living conditions or create a person with a serious medical condition, we usually also impose on these persons an increased risk to have overall bad lives. And this risk explains our obligations in non-identity cases. Henning’s proposal is insightful but, as I will argue, ultimately unsuccessful. I show that his view (i) relies on a dubious asymmetry between certain benefits and chances of benefit, thereby undermining a plausible ex ante understanding of risk, (ii) has implausible implications, and (iii) gives the intuitively wrong explanation of our obligations. On this basis, I suggest a revised understanding of the relevant risk-based objection. This results in a less ambitious solution to the non-identity problem but an overall more plausible view.