The role of intuition and deliberative thinking in ...

1 Brief articleThe role of intuition and deliberative thinking in experts superiortactical decision-makingJerad H. Moxleya, , K. Anders Ericssona, Neil Charnessa, Ralf T. KrampebaDepartment of Psychology, Florida State University, Tallahassee, FL 32306, USAbDepartment of Psychology, University of Leuven, Leuven, Belgiumarticle infoArticle history:Received 6 May 2011 Revised 11 March 2012 Accepted 14 March 2012 Available online 26 April 2012 Keywords:ChessDeliberate practiceSearchIntuitionDual process theoriesExpertiseDecision makingProblem solvingabstractCurrent theories argue that human decision making is largely based on quick, automatic,and intuitive processes that are occasionally supplemented by slow controlled delibera-tion.

2 Researchers, therefore, predominantly studied the heuristics of the automatic systemin everyday decision making. Our study examines the role of slow deliberation for expertswho exhibit superior decision-making outcomes in tactical chess problems with clear bestmoves. Our study uses advanced computer software to measure the objective value ofactions preferred at the start versus the conclusion of decision making. It finds that bothexperts and less skilled individuals benefit significantly from extra deliberation regardlessof whether the problem is easy or difficult.

3 Our findings have important implications forthe role of training for increasing decision making accuracy in many domains of expertise. 2012 Elsevier All rights IntroductionIn his Nobel prize lecture Daniel Kahneman (Kahneman,2002), described human decision making using two inter-acting systems. System 1 ( intuition ) is fast, automatic,and effortless, while System 2 ( deliberative thinking ) isslow, controlled, and effortful. In this model System 1quickly proposes solutions, and System 2 monitors System1, helps solve problems where the answer is not readilyapparent, and attempts to monitor and correct any biasesof System 1 (Kahneman & Frederick, 2005).

4 Most recent re-search has examined biases of System 1 for simple judg-ments (Gilovich, Griffin, & Kahneman, 2002; Tversky &Kahneman, 1974), but many scientists argue that heuris-tics and intuitions lead to better and more satisfying deci-sions in everyday life (Gigerenzer & Brighton, 2009) andcharacterize even the decision making of experts (Dreyfus& Dreyfus, 1986; Kahneman & Frederick, 2005; Klein,1998). The present study collected data on how the accu-racy of generated decision options changes from the startto the finish of the decision making process as a functionof the individual s level of expertise and the difficulty ofgenerating objectively superior Skilled decision makingExpertise researchers (Benner & Tanner, 1987; Dreyfus& Dreyfus, 1986) have argued that experts decisions areprimarily the result of accumulation of extended experi-ence, leading to fast intuitive decisions.

5 However, objectiveaccuracy of expert decisions has been found to be generallylow and in many domains of expertise longer professionalexperience has not been associated with better decisions(Choudhry, Fletcher, & Soumerai, 2005; Ericsson, 2007;Swets, Dawes, & Monahan, 2000; Tetlock, 2005). In do-mains, such as athletics, typing, and music, where we seereproducibly superior performance, expertise is not0010-0277/$ - see front matter 2012 Elsevier All rights Corresponding Moxley).Cognition 124 (2012) 72 78 Contents lists available atSciVerse ScienceDirectCognitionjournal homepage: with length of experience but with the durationof training activities (deliberate practice) that give imme-diate accurate feedback and opportunities for gradualimprovement (Ericsson, Krampe, & Tesch-R mer, 1993).

6 One of the best examples is chess where deliberatepractice such as the reported solitary studying of grand-master games has been found to be closely related to chessratings based on tournament success which is related toplayers ability to decide the best chess move in the labora-tory (de Groot, 1978; Gobet & Charness, 2006). While accu-rate feedback has been argued as necessary for thedevelopment of superior intuition (Hogarth, 2008) the lo-cus of individual differences in chess skill has not yet beenclearly linked to either System 1 or System 2 theories of chess expertise (Chase & Simon,1973;Gobet & Simon, 1996a) hypothesize that intuitiveprocesses based on pattern recognition, allow experts torapidly retrieve the selected action from memory.

7 In thesetheories the acquisition of a large body of knowledgelinked to familiar patterns allows the expert to recognizepatterns which cues the retrieval of the best this hypothesis, major theorists (Dreyfus &Dreyfus, 1986; Kahneman & Frederick, 2005; Klein, 1998)have described chess as an example of a domain whereexperts rely on superior intuition as stated inDreyfusand Dreyfus (1986, p. 36) Serious tournament chess in-volves deep deliberation, of movechoice depends surprisingly little on anything beyond pureintuitive response.

8 While the reliance on intuition does notpreclude a role for deliberation, intuition is thought to bemore important . Past research on deliberation in chessSincede Groot (1978)amount of search has been quan-tified by counting the number of different chess moves ormove combinations that a chess player verbalizes whileselecting the best move for a chess position. Although bet-ter chess players consider more alternative chess moves toa deeper level (Bilalic , McLeod, & Gobet, 2008; Charness,1981), previous analyses have not been able to determinewhether additional search increases the quality of moveselection by experts or mainly confirms the expert s ini-tially generated intuitions (Dreyfus & Dreyfus, 1986).

9 Research has provided some empirical support for theidea that experts gain less from See for instanceGobet and Simon (1998). For exam-ple, studies analyzing expert games (Calderwood, Klein, &Crandall, 1988) as well as tournament outcomes (Burns,2004) have argued that chess experts are less affected bytime reductions for move selection and thus by inferencewould benefit less by the time-consuming deliberation. Onthe other hand, other studies have found that rapid chessplay is associated with decrease in chess performance, butless than might be expected (Gobet & Simon, 1996b) andis associated with selection of worse chess moves (Chabris& Hearst, 2003).

10 The present studyThis study will approach the question differently by col-lecting verbal protocols during the decision making pro-cess and using controlled laboratory tasks. We used achess computer program, which selects better moves fortactical chess positions than the best human players, to as-sess the quality of chess moves on a scale that approxi-mates measurement on the interval level. This allows usto quantify the differences in selected move quality atthe start and at the end of the decision making processand to see how much experts and skilled tournament chessplayers benefit from additional time to search and deliber-ate.