Drawing upon the historical relationship between statistics, probabilistic reasoning and life insurance, the article argues that mathematical calculation played a necessary but limited role in making markets for life insurance. Insuring publics have been fairly consistently cautious in the use of probabilistic and statistical reasoning to inform investment in life insurance. In this they follow a pattern set by early insurance companies who themselves were slow to alter their commercial practices in line with emerging knowledge. I examine some of the reasons for this glacial pace and some of the ambiguities on which statistical 'certainties' were built as part of an argument that the role of statistics and mathematics in market calculation is both less and more than it seems. This is manifest in the history of industrial life assurance, an industry with a phenomenally successful track record in the mass enrolment of consumers. Unlike their predecessors, industrial companies disdained swamping their target markets with probabilistic arguments in favour of a very different sort of argument that, nevertheless, carried a trace of statistical thinking with it. This trace came in the form of 'good, average men', the agents who became industrial insurance's core marketing device and who translated the essentials of a statistically informed product into a more palatable, more calculable form.