Addendum: My instincts point towards rule consequentialism - as a finite limited-information machine, optimizing my adaptations seems to me like generally a good way to win the better outcomes.