Many of the programs ideals have parallels in the discussion at MeatBall, though vocabulary is different:
A phenomenon of Outward Bound is that teamwork tends to deteriorate on the last day or two, a manifestation of the IteratedPrisonersDilemma when N (the number of games) is known in advance. As all players know their interaction will end in two days, the incentive to cooperate is greatly reduced. Note that according to the Backwards Induction Argument (which is based on this observation) rational players should not even cooperate on the first move.
N is not measured in days, but rather in interactions. There are generally 1-3 meaningful interactions per day, perhaps more. Towards the end of the relatively brief OutwardBound experience, it becomes clear that there are a limited number of interactions left, but this is not clear, say, five days before the end. The uncertainty in N makes the IPD work, and I suppose there is a mathematical model that could describe this quantitatively and identify the amount of uncertainty necessary. I have never claimed to be a mathematician, so I will (heh) leave the proof of this as an excercise to the reader.