Along with Ivan Bratko and Peter Tancig in the early 80s, Simona Tancig researched on cognitive processes in the game of chess and on detection of positional patterns and chunks. Similar to the attempts of Adriaan de Groot[4], and William Chase and Herbert Simon[5][6], they conducted experiments where chess players had to reconstruct complicated middlegamepositions after a short view on diagrams, with the novel aspects of collective reconstruction, as published in the ICCA Journal[7] and Advances in Computer Chess 4. 24 positions were chosen from a chess magazine with 28 pieces on average. The positions were of four types according to the opening from which the position arose (6 each Ruy Lopez, Sicilian, King's Indian and Queen's Indian). 20 players in two groups were the subjects, one group rated from 2300 to over 2500, the second from 1800 to 2100.
Reconstruction Factor
The success of reconstruction was measured by the so-called reconstruction factor:
where P is the set of pieces (including the square information) in the original, and R is the set of pieces in the reconstructed position, with 0 <= F <= 1. Using a standard bitboard board-definition with arrays of 12 bitboards for each piece-type and color, the reconstruction factor could be calculated with following routine aggregating population counts of intersections and unions, and the final double floating point division:
double reconstructionFactor(const BitBoard* P, const BitBoard* R){int i =0, u =0;for(int pt =0; pt <12; pt++){
i += popcount(P[pt]& R[pt]);
u += popcount(P[pt]| R[pt]);}return(i == u)?1.0:((double) i /(double) u);/* considers i == u == 0 */}
The measure is different from those used by De Groot, Chase and Simon. Its advantage is that it not only rewards correctly recalled pieces, but it also penalizes incorrectly added pieces, where the previous used measure was not sensitive of.
Individual Results
Following individual reconstruction results were determined for the two groups:
Simona Tancig (2009). Expert Team Decision Making and Problem Solving: Development and Learning. in Interdisciplinary Description of Complex Systems, Vol. 7, No. 2, pdf
^Adriaan de Groot (1965, 1978). Thought and Choice in Chess. Mouton & Co Publishers, The Hague, The Netherlands. ISBN 90-279-7914-6, amazon, google
^William Chase, Herbert Simon (1973). The Mind’s Eye in Chess. Visual Information Processing: Proceedings of the Eighth Annual Carnegie Psychology Symposium (ed. W. G. Chase), pp. 215-281. Academic Press, New York. Reprinted (1988) in Readings in Cognitive Science (ed. A.M. Collins). Morgan Kaufmann, San Mateo, CA.
a Slovenian psychologist, since 1991 associate professor of developmental psychology at Faculty of Education of University of Ljubljana, where she also received her Ph.D. in psychology in 1983 with the thesis title Some Cognitive Processes in the Game of Chess [1]. She was the main researcher and collaborator at multiple research projects in the areas of cognition, decision making, learning, education, psychomotor development, and mathematical cognition. Her current research interests are in metacognition, self-regulation, collaborative learning and expert knowledge, embodiment and empathy [2]. In 2010, Simona Tancig received the Award of the Republic of Slovenia in the field of education.
Table of Contents
Chess experiments
Along with Ivan Bratko and Peter Tancig in the early 80s, Simona Tancig researched on cognitive processes in the game of chess and on detection of positional patterns and chunks. Similar to the attempts of Adriaan de Groot [4], and William Chase and Herbert Simon [5] [6], they conducted experiments where chess players had to reconstruct complicated middlegame positions after a short view on diagrams, with the novel aspects of collective reconstruction, as published in the ICCA Journal [7] and Advances in Computer Chess 4. 24 positions were chosen from a chess magazine with 28 pieces on average. The positions were of four types according to the opening from which the position arose (6 each Ruy Lopez, Sicilian, King's Indian and Queen's Indian). 20 players in two groups were the subjects, one group rated from 2300 to over 2500, the second from 1800 to 2100.Reconstruction Factor
The success of reconstruction was measured by the so-called reconstruction factor:where P is the set of pieces (including the square information) in the original, and R is the set of pieces in the reconstructed position, with 0 <= F <= 1. Using a standard bitboard board-definition with arrays of 12 bitboards for each piece-type and color, the reconstruction factor could be calculated with following routine aggregating population counts of intersections and unions, and the final double floating point division:
The measure is different from those used by De Groot, Chase and Simon. Its advantage is that it not only rewards correctly recalled pieces, but it also penalizes incorrectly added pieces, where the previous used measure was not sensitive of.
Individual Results
Following individual reconstruction results were determined for the two groups:For each piece type the table indicates that pawns play a major role in human positional perception:
See also
Selected Publications
External Links
References
What links here?
Up one level