I am making a tautological
revealed preference argument as follows: people do stuff they like and we know they like to do that stuff cause that is the stuff they do. Therefore everyone is "alike" in the sense that they do the stuff they like to do.
This is not a very interesting sense in which people are all alike. What is interesting is their very varied preferences, and hence their very varied utility functions. In this sense, which I suggest is more interesting, people are often very different from one another.
And as others have pointed out, your "like to do" is ambiguous as between a thing (i) that they gain pleasure from doing, (ii) that they satisfy a preference by doing, and (iii) that is the rational thing to do consistent with their overall preference structure. For those with other-regarding preferences, (i) and (ii) can come apart. For those with higher-order preferences not to act on their lower order preferences, (ii) and (iii) can come apart (eg Ulysses and the sirens). And which preference will be revealed by a person's behaviour depends upon further complexities such as the possibility of acting out of weakness of will. (Hence to understand Ulysses's overall preferences we have to notice his instruction to his men to tie him to the mast, rather than his cries for release when the sirens come within earshot. Similarly for the knowing lycanthrope who instructs his allies to wrap him in chains of silver and then ignore his entreaties later in the evening. )
What I intended was probably closer to the (tautological) concept of utility maximization.
If I understood it right, your initial post suggests that everyone has a strong self-regarding survivalist preference, but that different beliefs about (i) the likelihood of social reciprocation, and (ii) the usefulness of departing from expected ways of doing things, will produce different behaviours that are expressive of the different alignments.
If I wanted to explain why some people are more generous than others, though, I wouldn't confine myself to positing different beliefs about the likelihood of reciprocation. Equally for those who act viciously. I'd also ask myself whether or not they have genuine other-regarding preferences, whether altruistic or hostile.
At least as I've tended to understand it, I think that D&D "good" is meant to describe those who actually have, and who act upon, altruistic preferences. The definition of D&D "evil" has always been a bit ambiguous as between other-regard and self-regard, but on at least some readings it seems to be used to refer to those who have, and who act upon, other-regarding but non-altruistic preferences (ie a desire that others suffer).
I've always assumed that, in D&D terms, a person who acted in a generous way but purely out of a self-regarding expectation that such behaviour will be rewarded is "neutral" as between good and evil, although the sort of neutral that you wouldn't necessarily mind having as a next-door neighbour. Conversely, there has always been some ambiguity as to whether a person who acts in a hostile fashion, but purely out of a self-interested desire to ensure their own survival, is neutral or evil. When played as a PC, this person tends to be given the benefit of the doubt, and hence neutral status. For NPCs, it's a bit more ambiguous (for example 1st ed Oriental Adventures, if I remember correctly, suggests that greedy merchants are neutral evil).
And this doesn't even get to the issue of whether a method of analysis that was invented to explain the behaviour of modern individuals in a capitalist market (ie modern economic theory) is really exportable to what are ostensibly very different social and political contexts (ie the D&D gameworlds, which however varied are rarely meant to be modern capitalist societies).