Evidence for Chunking Theory

Chess master's eye movements
(link to source)

One of the main pieces of evidence supporting the chunking theory is the chess experiment carried out by De Groot (1965). In this pretty simple experiment, De Groot was able to show the difference in the way that experts and novices reconstruct certain aspects of chess in their minds.

The Experiment

In the experiment, De Groot took a chess position from a master game. The subjects, possessing different levels of chess knowledge, were not aware that the position was taken from a master game. De Groot then presented the subjects with the position for a certain amount of time-something between 2 to 15 seconds. After the presentation of the position, it was removed from their sight and the task for the subjects was to reconstruct the position they had just been presented with using a different board. The ability of a subject's memory would be based on the number of pieces that they could correctly place on the new board.

The Findings

What did De Groot find? He found that his grandmaster was able to remember almost every piece of the presented position perfectly-about 93% correct-while his weakest subject could barely place 50% of the pieces correctly on the new board. Based on these findings, De Groot stated that masters of chess do not encode the position as isolated pieces. Rather, the chess masters see the positions as "large, mostly dynamic 'complexes.'" (Gobet, 1998) In addition, the complexes the chess masters encode also incorporate the empty squares, because the empty places play an important role in reconstructing the exact position. The ability of the masters to encode these large complexes is based on the amount of knowledge and experience that they have acquired over time during both their study and practice of chess.

However,as mentioned before in the chunking theory, section, the re-evaluation of De Groot's experiment by Chase and Simon is what actually led to the manifestation of the chunking theory. After they studied the time differences that occurred during the placement and recall task of the chess experiment, they found that the master recalled bigger chunks and more chunks. They also found that the master's stored a larger number of patterns, once again attributing the theory being based on pattern recognition in long-term memory (LTM).

The Figure Explained As A Direct Application To The Chunking Theory and The EPAM Model

The figure is a schematic representation of MAPP, a computer model. This model incorporates both concepts of chunking theory and the simulated memory process of chess players. The computational model shows the same concepts discussed with regard to the EPAM model. The upper part of the model shows the learning phase that goes through the EPAM model discrimination net. The bottom section of the figure shows what happens during a recall task and how the pieces are fed through the discrimination net and when the chunk is recognized a symbol is output and the symbols are placed in short-term memory (STM). After this, the masters reconstruct the position using the symbols in STM and the chunks that the symbols point to in LTM.

The figure not only shows how the chess experiment by De Groot shows the basis of the chunking theory, but also how the computational model composed after the re-evaluation of De Groot's experiment is supportive of the chunking theory while explaining the basic components of the EPAM Model.

Please continue exploring the site!

(Gobet, 1998)