Size-selective predation has been proposed to be one important evolutionary force shaping life-history traits in guppies (Poecilia reticulata). Populations living in the presence of the ring-tailed pike cichlid (Crenicichlasaxatilis) are smaller, mature earlier, allocate more energy to offspring and get more and smaller young than guppies in localities without Crenicichla. We investigated if Crenicichlasaxatilis is a size-selective predator, if the selectivity is a result of active choice and if the optimal prey size can be explained according to an optimal foraging model. In single-prey experiments we quantified the predators' pre-capture costs (time), capture success, and post-capture costs (time) for four different prey sizes spanning from 10 to 40 mm total length. To see which of the components of the prey cycle the predator takes into account for its choice, we then predicted prey values and optimal prey size with 6 different models that included one or more of the prey cycle components. In two multiple prey experiments, the cichlids were given the choice of the two and four different prey sizes simultaneously. Crenicichlasaxatilis actively selected the largest guppies in both cases. The three prey-value functions that included handling time (post-capture cost) did not accurately predict the prey choice. Instead the prey-value functions that took into account pre-capture cost (approach and attack time) were able to correctly predict the choice of the largest guppy size, suggesting that pre-capture costs may be more important than post-capture costs for prey choice in Crenicichlasaxatilis. The study confirms that Crenicichlasaxatilis is a size-selective predator selecting large guppies, while earlier evidence for selectivity for large prey in Crenicichla cichlids has been weak and equivocal. Our result strengthen the possibility that size-selective predation is a mechanism in life-history evolution in guppies.