A Behavioral Investigation of the FlipIt Game

Abstract: Security decisions are rarely made at a singular point in time. Though many models evaluate choice under uncertainty with regards to many options, fewer address the problem of when to act. The FlipIt game captures this temporal choice, only allowing players to choose when, and not how, to act. Decisions of this sort are encountered by managers choosing when to address fraud, computer users selecting when to update their software, and consumers deciding when to check their credit score. Recent investigations analyze the FlipIt game from a theoretical perspective, but the question remains- how do people actually act when given temporal decisions? To answer this question, we conduct a behavioral investigation of the FlipIt game through a Mechanical Turk experiment with over 300 participants. In our study, each participant is matched with a computerized opponent in several fast-paced rounds of the FlipIt game. We find that participant performance improves over time (however, older participants improve less than younger ones). Further, there are significant performance differences with regards to gender and an individual difference variable reflecting the extent to which individuals are inclined towards effortful cognitive activities (i.e., the need for cognition). We further vary the amount of information that participants have available about the actions of the computerized opponent with six different experimental treatments and find significant statistical effects