Evolutionary psychologists believe that there must be an important benefit of play, since there are so many reasons to avoid it. Animals are often injured during play, become distracted from predators, and expend valuable energy. In rare cases, play has even been observed between different species that are natural enemies such as a polar bear and a dog.[9] It has also been noted that play seems to be higher up on a hierarchy of needs. For example, stressed and starving animals do not play (making it a broaden and build behavior).[10]

One theory – Play as preparation – was inspired by the observation that play often mimics adult themes of survival. Predators such as lions and bears play by chasing, pouncing, pawing, wrestling, and biting, as they learn to stalk and kill prey. Prey animals such as deer and zebras play by running and leaping as they acquire speed and agility. Hoofed mammals also practice kicking their hind legs to learn to ward off attacks. While mimicking adult behavior, attacking actions such as kicking and biting are not completely fulfilled so that they won't injure each other. In social animals, playing might also help to establish dominance rankings among the young to avoid conflicts as adults. On the other hand, this view runs into some problems; the behaviors practiced are often quite different, or even exactly the opposite of those required in the equivalent real life situation.[10]

Researcher John Byers describes how the amount of time spent at play for many mammals (e.g. rats and cats) peaks around puberty, and then drops off. This corresponds to the development of their cerebellum, suggesting that play is not so much about practicing the exact behaviors, as much as building general connections in the brain. Research by Sergio Pellis and colleagues discovered that play may shape the brain in other ways. Young mammals have an overabundance of brain cells in their cerebrum (the outer areas of the brain – part of what distinguishes mammals). Play has been evidenced to help the brain clean up this excess of cells, resulting in the more effective cerebrum of maturity.[10]

Marc Bekoff describes a Flexibility Hypothesis which attempts to incorporate these newer neurological findings. It argues that play helps animals learn to switch and improvise all behaviors more effectively. Animal researcher Marek Spinka believes that playing helps animals learn to handle new and surprising events.[11] There may, however, be other ways to acquire even these benefits of play – the concept of equifinality. The idea is that the social benefits of play for many animals, for example, could instead be garnered by grooming. Patrick Bateson maintains that equifinality is exactly what play teaches. In accordance with the flexibility hypothesis, play may teach animals to avoid "false endpoints." In other words, they will harness the childlike tendency to keep playing with something that works "well enough," eventually allowing them to come up with something that might work better, if only in some situations. This also allows mammals to build up various skills that could come in handy in entirely novel situations.[10]

Peter Smith warns against a "play ethos." He says we must keep things in perspective, and let real evidence – rather than wishful thinking – guide our beliefs about play.[10]