Evolution produces good and bad people

I discovered in http://plus.maths.org/latestnews/sep-dec08/gametheory/index.html a rather nice explanation of how evolution can bring about a balance in society between good and bad people. Society can have some trusting and some suspicious "Alices", and some trustworthy and some deceitful "Bobs". Encounters between the four possible combinations of Alice and Bob have different payoffs. The best payoff for Bob comes from being deceitful, which means that the best strategy for Alice is to be suspicious. But the best payoff for Alice and Bob combined comes from being trusting and trustworthy.

If people are simply one or the other, evolution would work according to the old "survival of the fittest" myth and produce suspicious and deceitful people. Mutations producing trusting and trustworthy people might produce a locally successful society but it would be unstable because deceitful individuals would still do better.

But then suppose we have social awareness, whereby individuals are able to observe the behaviour of others in previous social interactions and modify their strategies accordingly. It is better to trust a trustworthy person and it is better to be trustworthy so as not to be distrusted. Evolution will begin to produce more "good" people, or at least people who are good more often. But in the long term this will only work if there is the possibility of deceit and suspicion to keep the need for social awareness alive, and this requires the reality of at least a few "bad" people. So a stable society with the best overall outcome for all will have some bad people in it.

This does not rule out the possibility that matters can be improved by culture. Training people in social awareness can reduce the requirement for people to be tested by real-world deceitful or suspicious encounters.