The "problem" of altruism - at least in terms of evolutionary theory - was obvious even to Charles Darwin.
Writing in The Descent of Man in 1871 Darwin acknowledged that, "he who was ready to sacrifice his life, as many a savage has been, rather than betray his comrades, would often leave no offspring to inherit his noble nature."
It's a puzzle because the logic of natural selection leads us to expect animals to behave in ways that increase their chances of survival and reproductive success - to act selfishly.
And yet altruistic behaviour is common: when a bee stings it sacrifices itself for the good of the colony; vampire bats regularly regurgitate blood for others in the group; and Vervet monkeys scream out warnings of an approaching predator even though the calls give their own position away.
Darwin's own, somewhat tentative theory, was that altruism may have evolved as a result of inter-group rivalry. A tribe that included individuals who were willing to sacrifice themselves for the common good should triumph over its more selfish neighbours, allowing its genes - including those for altruism - to prosper over time.
But the problem with group selection is that it's vulnerable to subversion from within - the Vervet monkey that keeps quiet benefits from the warnings of others while remaining hidden. Genes for Vervet sneakiness should flourish in subsequent generations.
One way evolutionary biologists test their ideas about altruism, is by getting people to play carefully designed economic games in the lab. These experiments show that humans cooperate much more than naked self-interest demands - even when players are trying to maximise their position.
Proof then that humans are altruistic, preferring to cooperate and for the group to succeed.
Well not according to the Oxford biologist Professor Stuart West, who argues in a paper published in the Proceedings of the National Academy of Sciences, that humans are actually no more altruistic than bacteria - it's just that we're really bad at playing games.
"How people behave in economic games, where they can choose to be selfish or cooperative, can be explained more easily by 'mistakes' than wanting others to succeed."
The experiment involved 168 people in a series of four games where they were able to decide how much money to contribute to a public project. Each of the games was designed to reward a different outcome - from fully altruistic to completely selfish.
But the results showed that even when an extreme strategy delivered the best results, competitors still hedged their bets, refusing to commit all their money.
Professor West believes that could be because humans are programmed to mistrust exposed positions, or it could be because we're just very bad at assessing the consequences of our actions.
"This could derive from a psychology that avoids extreme behaviours, which could be very costly if they go wrong, or indicate that the sort of simple everyday rules of thumb we use to make these judgements misfire".
Either way it's not altruism.