The Extensive Study Of Chunk Mechanism And Its Relation To Chess Skill

Understanding and explaining expert performance in special domains is a traditional field of cognitive psychology. As a general cognitive theory, chunking theory has made quantitive and effective predictions and explanations of expert performance on perception, memory, problem solving and so on. Chunk theory proposes that experts can encode more information into more and larger complex chunks while novices can not. Recently, however, there are researches questioned the hypothesis of chunking theory and the nature and function of perceptual chunks. These questions mainly focus on three points:First, how expert and novices encode the information on stimuli positions, on basic information or on abstract information; Second, whether perceptual chunks can differentiate individuals from different skill levels; Third, whether chunks can help chess players recall information related to chess move selection. Besides, although current researches identified the expert performance as well as chunks and their structural characteristics in labs, no work has been done to account for expert performance by using chunking mechanism they identified.With respect to the questions and the deficiency of the current researches, this study for the first time employed the expert performance approach in chess field, and extensively analyzed the structural and especially the content characteristics of chunks. This study then further discussed the functional relationship between chunking mechanism and expert performance, and evaluated the corresponding hypothesis for chunking theory.In chess, three experiments were designed and carried out. The first experiment studied chess experts and masters’ short-term memory storage and the function relationship between chunk characteristics and presentation time. Compared to Gobet and Simon (2000), this study manipulated the factor of age and only used young participants. More importantly, this experiment take a further step and looked into how chunking mechanism mediate expert performance and players’chess skill. To make a comparison to experiment one, the second experiment tested novices’ short-term memory storage and the relation between the number of presentation times and chunk characteristics. Those two experiments showed that experts’ STM capacity was 3 chunks and the largest chunk contained 10-12 pieces, while novices recalled 1 chunk with 3 pieces included. For experts, on game positions, the number of chunks was negatively correlated with presentation time, while the size of the largest chunks correlated with presentation time positively. Both characteristics of chunks correlated with presentation time positively on random positions. For novices, both chunk characteristics correlated with number of presentations positively on each position type. In addition, chunk mechanism effectively explained and mediated memory performance and chess skill.The third experiment was carried out with the aim to identify specific chunks for individual players. This experiment used an extensive Chase-Simon paradigm and a wider range of participants both on age and chess skill. Not only the structural characteristics of chunks, the content characteristics of chunks were explored as well. Experiment 3 compared two independent chunk-identification techniques, namely repeated-recall chunks which relies on the recall orders of chess pieces and response time technique which relies on the latencies between chess pieces. Consistent significant differences between skilled players and non-chess players on these characteristics were found on game positions. It is proved that both techniques can be used to identify chunks. Correspondingly, repeated-recall chunks correlated with response time chunks on both number and size of chunks. This study therefore concluded that researchers can choose the corresponding technique according to the experiment paradigm and the participant sample. To be specific, researchers can use repeated-recall technique when participants are recruited from a wide range of age and the stimuli were presented and recalled repeatedly. Researchers can use response time technique when participants are mainly young and the stimuli only presented for one time like most current studies did.To summarize, the important findings of this study are as follow:(1) Only after the representation is build up on basic elements (short time frame), abstractions and analogies may start to occur (longer time frame).(2) Both the structural characteristics and the content characteristics can differentiate players from different skill levels. Experts recalled more chunks, larger largest chunks, more semantic information in chunks, and the recall of orders of pieces within chunks were more flexible. It is therefore concluded that chunks can reflect the individual differences related to skill levels.(3) Skilled players recalled more pieces from the strategic area of the positions. They recalled more positions with both kings were recalled correctly and they recalled more semantic relations. This study therefore for the first time concluded that these evidence provide strong support to chunking theory’s hypothesis that stronger chess players recalled more information related to move selection.(4) Chunking mechanism is effective and can account for the total variance for expert performance and chess skill significantly. Besides, the size of the largest chunks explained the most part of variance for memory performance and chess skill. That is, the size of the largest chunks is the main indicator to players’chess skill.To summarize, this paper responded effectively to the current questions about chunking theory and the nature of perceptual chunks, therefore this paper provide empirical evidence to the corresponding hypothesis for chunking theory. The findings and conclusions extended and perfected chunking theory. Researchers can better understand expert performance and skill development from the view of chunking.