In 1736/37, Joseph-Nicolas Delisle and Jean Jacques Dortous de Mairan communicated about the clocks that would enable the astronomers of the Saint Petersburg observatory to make highly exact observations. Delisle, who was in charge of the Saint Petersburg observatory, demanded old-fashioned clocks in the manner of Huygens. Mairan, well-versed in astronomy himself, recommended equation clocks. The article uses these seemingly inappropriate preferences to discuss eighteenth-century notions of accuracy and precision in clocks. It analyses the multiple factors that influenced expectations regarding the performance of timekeeping instruments, and draws attention to handling and monitoring practices. The latter reflected the individual user's purposes and experience, but also affected the clocks' going. Furthermore, the article presents the result of a statistical analysis, which serves to evaluate the historical performance of the Saint Petersburg observatory clocks and provides a foil against which Delisle's judgement of them is examined.