This concept hybridizes the Prisoner's Dilemma with the Milgram experiment. In the Prisoner's dilemma, a good outcome is achieved when both prisoners cooperate, but one prisoner can achieve a better outcome for himself at the expense of the other by not cooperating. (link)

The Milgram experiment
involved an authority instructing the participant to give shocks to an unseen person.

In Prisoners Dilemma Solitaire the participant plays a game of chance via computer with another participant. If the participant wins he gets a small sum, plus the right to administer a shock to the other participant. The other participant may opt to quit the game before receiving the shock, and then the remaining participant wins and with it wins a large sum. The voltage of possible shocks is chosen by the one administering, and the maximum possible increases as the game goes on.

The twist is that there is only one participant. On losing a round, he receives a shock exactly equal to the one he gave last.

Will participants surmise that the best approach is not give shocks at all and rather play and accumulate small sums indefinitely? Or will a participant escalate and escalate shocks, in effect shocking himself until he is compelled to quit and lose?