Infanticide is hypothesized to be an adaptive strategy in many mammalian taxa. While most authors focus on infanticide by new male immigrants, natural selection should favor infanticide under a wider range of conditions, including sometimes by potential fathers who could increase their paternity probability in a new offspring. Additionally, most authors have focused on direct male protection of infants as a counterstrategy to infanticide, while not considering voluntary sharing of paternity to reduce infanticide by other males. Here I model males' decisions about whether to commit infanticide and explore how infanticide risk affects optimal male and female mating strategies. When males cannot easily and effectively protect infants, infanticide risk creates a fitness landscape with two adaptive peaks, one representing complete paternity certainty and the other which represents a compromise between maximizing paternity and minimizing infanticide risk. Which of these adaptive peaks represents the fitness-maximizing global optimum for a given male depends on a population's socioecology and characteristics of the male. While high ability to protect infants, high rank stability, low extra-group takeover possibility and high male-male relatedness within a group encourage males to maximize paternity, in many ecological contexts, males may adaptively reduce their paternity probability to reduce the risk of infanticide. In particular, populations with relatively long lactation periods (compared to gestation periods), frequent rank reversals, a lack of male philopatry, frequent takeovers by extra-group males and limited ability of males to protect infants (e.g. because of a fission-fusion or solitary social organization or coalitions formed among subordinate males) all increase the probability that a male will maximize his reproductive success by voluntarily giving up a significant portion of his potential paternity probability. Nevertheless, females should tend to more closely agree with the optimal paternity probability sought by subordinate males, and thus may seek extra-pair copulations beyond the amount optimal for the a-male, fueling inter-sexual competition. Thus, explicit, mathematical consideration of the fitness effects of male and female mating decisions, including their effects on infanticide probabilities, enhances our understanding of the dynamics of mammalian intrasexual and intersexual competition in a number of ways.