The Morra game as testbed for automatic and controlled cognitive processing in an ecological setting

Morra is a 3000-year-old hand game of prediction and numbers. The two players reveal their hand simultaneously, presenting a number of fingers between 1 and 5, while calling out a number between 2 and 10. Any player who successfully guesses the summation of fingers revealed by both players scores a point. The game is extremely fastpaced, making it very difficult for players to achieve a conscious control of their game strategies. In this study, we used Morra as a naturalistic setting to investigate a number of cognitive skills and their interaction. We analyze the abilities of naı¨ve, partially trained, and expert players of Morra to generate seemingly random numeric sequences to predict their opponent’s numeric sequences and tested our participants with a battery of neuropsychological tests including measures of working memory, executive functions, processing speed, and attention. Finally, sequence prediction and random number generation performances were compared and modelled against the optimal strategies of a non-deterministic artificial intelligence specifically built to play Morra. Preliminary results indicate that Morra skills (a) do not correlate with the proposed neuropsychological measures; (b) are mostly automatic; (c) are the result of the interaction between perceptual, cognitive and motor components. Results are discussed in the framework of the dual processing theory and its automatic-parallel-fast versus controlled-sequential-slow polarities: in spite of the extreme speed of the game and of the lack of awareness of their strategies, expert players have a better statistical prediction of the opponents’ sequences and generate less predictable sequences than naı¨ve players.