Use of a Prediction Paradigm to Evaluate Proficient Decision Making

Abstract

The contrast between proficient and novice chess players was examined. A chess game played between two experts was used as the target. Subjects were never shown the actual game but were asked to predict the moves played. Feedback was given after each prediction. The proficient players were significantly more accurate than novices. They also were more likely to generate the correct option as the first one they considered and made fewer guesses about the predicted move. The data are interpreted as supporting a recognitional model of decision making. In addition, the findings suggest the value of using the prediction paradigm to study decision making in ill-defined tasks where there are no criteria for correct responses. Implications for training are discussed.