A Regret-Based Model for Wishful Thinkers

Simone Duca (RUB & HHU) gives a talk at the MCMP Colloquium titled "A Regret-Based Model for Wishful Thinkers". Abstract: The study of decision making can be divided into three categories, according to how much information is available to the agent. We talk of decision making under risk, uncertainty and ignorance, whenever the agent is facing a decision under complete probabilistic knowledge, incomplete probabilistic knowledge and no probabilistic knowledge re- spectively. In my talk, I wish to focus on decision making under ig- norance and propose a decision model where the agent either has no prior probabilities at all or she deliberately ignores them when given. The model is an attempt to capture two facts: a) people are not very good at reasoning with probabilities and b) people tend to be unreal- istically optimistic when making a decision. In other words, in many situations, people are inveterate wishful thinkers. This can be repre- sented by applying a non-probabilistic rule which aims at minimising the worst-case scenario regret. I discuss some properties and applica- tions of the model and give evidence for the descriptive aptness of the model.