By measuring the actual number of risk acts engaged in by HIV-positive men participating in the Positive Health (PH) cohort study, this paper sets out to document the distribution of risk acts, to report on the proportion of acts of unprotected anal intercourse with casual (UAIC) partners that occurred between HIV-positive men (i.e. seroconcordant positive) and to examine the factors that differentiated men who 'frequently' compared with 'sometimes' or 'never' engaged in unsafe UAIC (i.e. UAIC with serononconcordant partners: partners who have not tested positive for HIV). The findings show that 42.6% of all UAIC acts occurred between seroconcordant HIV-positive partners, posing no risk of HIV infection to an HIV-negative person. A minority of participants (10%) accounted for the majority (70.7%) of the unsafe acts of UAIC. The HIV-positive men who 'sometimes' engaged in unsafe UAIC had higher treatment optimism scores and were more likely to use Viagra in comparison with those who did not engage in such risk. Those who reported 'frequent' engagement in unsafe UAIC were more likely to engage in a range of esoteric sexual practices, be slightly less well educated and be taking antiretroviral therapy compared with HIV-positive men who 'sometimes' engaged in unsafe UAIC. As such, taking ART but not viral load, predicted frequent unsafe UAIC. When considered alongside earlier studies, these results suggest that HIV-negative men who engage in esoteric sexual practices may be at increased risk of HIV transmission, not necessarily because they engage in esoteric sex practices but because of the sub-cultural milieu in which esoteric sex is occurring. The findings from this study also endorse the measurement of UAIC acts as a useful gauge of risk.