Does it takes more than practice and experience to become a chess master?
Many people believe that practice and experience are the key to success on the chessboard. But a recent study suggests that it takes more than that to become a chess master. In my latest blog post, I would like to share the findings of this study with you and discuss what it really takes to become a chess master.
I´ll examine the stories of a child prodigy and adult chess players to find out what made them masters of their craft. Join me on this fascinating journey into the world of chess and let's discover what it really takes to reach the top of this demanding game.
| Introduction |
Reaching master-level performance in chess requires at least 10 years or 10,000 hours of deliberate effort. The primary source of individual differences in chess skill is deliberate, according to Ericsson.
Examining whether deliberate play or other chess-related activities were the focus of two studies. Individual differences in chess expertise can be explained by experience and other factors. Expertise in chess may be contributed.
The first study examined the amount of time spent by a gifted and young chess player. CS (First letters of the first and last name of the gifted child) had studied on her own and participated in other chess-related activities. CS spent little time studying alone, and spent little time studying alone. Some time is spent on other chess-related activities. Regardless, she achieved a remarkable proficiency in chess. It's tough to reconcile accomplishment with the 10-year or 10,000-hour rule. Finally, CS performed exceptionally. A visual short-term recall test performed well.
Study 2 investigated factors contributing to the chess ratings of 77 Chess players who are adults. Studying on their own and participating in other chess-related activities were both time-consuming. Contrary to the theory of deliberate practice, other factors include General fluid intelligence, specific fluid intelligence, and specific crystallized intelligence.
All contributed substantially to the prediction of chess ratings even after controlling for practice and other chess related activities. These results support the idea that studying on your own and playing chess is beneficial. It is necessary but not sufficient for achieving a high level of chess performance.
| Explanation |
Psychologists have been studying the factors that contribute to expertise in various domains, such as chess. There is a debate about the importance of natural ability versus practice in determining individual differences in performance. Some argue that practice is the primary factor, with natural ability playing a minor role. Research has shown that practice is crucial in building up patterns of knowledge, such as chess patterns, in long-term memory.
Simon and Chase (1973) found that after 10,000 to 50,000 hours of practice, chess players can play at a master level by storing familiar patterns. They emphasized the role of practice in skill acquisition and argued that the organization of a master's repertoire of information takes thousands of hours to build up. Ericsson et al. (1993) also highlighted the importance of practice but focused on the role of coaches and tutors in skill acquisition.
While some studies suggest that expertise is mainly due to practice, there is evidence that natural ability may play a significant role. Child prodigies, such as Bobby Fischer, show exceptional talent and performance at a young age. Ericsson argues that deliberate practice is essential for becoming an expert, with nearly anyone able to reach an international level with enough practice.
Recent studies have shown that the amount of practice, whether individual or group, predicts chess skill. However, there is variability in the amount of practice needed to reach master level, suggesting that domain-specific practice is necessary but not sufficient for expertise. Critics have questioned the methods used in studies supporting the importance of practice and have suggested that natural ability may play a more significant role than previously thought.
Overall, the debate continues about the relative importance of practice and natural ability in the development of expertise in various domains, including chess. While practice is crucial for skill acquisition, there is evidence that natural ability may play a role, especially in the case of prodigies. More research is needed to better understand the complex interplay between practice and innate talent in the development of expertise.
| Study 1 |
In Study 1, researchers conducted an investigation of a talented young chess player referred to as CS. The main focus of the study was to understand the development of CS's expertise in chess.
The researchers were particularly interested in tracking the trajectory of CS's chess skills and determining whether her dedication to studying and playing chess aligned with the commonly held belief that expertise in chess is primarily built through practice, especially deliberate practice.
In addition to analyzing CS's chess skills, the researchers also assessed her cognitive abilities through various tasks. These tasks were chosen based on previous studies conducted on cognitive abilities in relation to expertise in different fields. One notable study mentioned was Baumgarten's (1939) investigation of Sammy Reshevsky, a young chess player, recognized for his exceptional visual short-term memory skills.
Although Reshevsky performed poorly on most tasks, his outstanding performance on memory tasks stood out, showing his remarkable abilities. Similarly, Ruthsatz and Detterman's (2003) case study of a musical prodigy highlighted the significance of extraordinary short-term memory in exceptional talent. This emphasizes the role of specific cognitive abilities in contributing to expertise in various domains.
Overall, the study aimed to uncover the unique cognitive profile of CS, shedding light on the underlying factors that contribute to expertise in chess and other domains.
The Process
Participants
Young Chess Player
At the time of testing, CS was a 10-year-old female chess player who had obtained a United States Chess Federation (USCF) rating of 2141 (at the 96.6th percentile of the entire USCF population and at the 99th percentile of the entire USCF female population based on the database in USCF in 2015: 64,069 members and 8,982 female members, respectively).
In addition, CS was in the top 46 USCF females (regardless of residence or federation), and in the top 30 USCF females in the U.S.
Children
A total of 34 healthy 10-year-olds children were recruited from different elementary schools from Houston, Texas, and Taipei City in Taiwan. 16 of the children are boys and 18 are girls.
Adult Chess Players
A total of 79 chess players were recruited from different chess clubs and chess tournaments from Dallas, Fort Worth, College Station, Beaumont, Galveston, Houston, and surrounding areas in Texas and from Taipei City in Taiwan.
Two participants were excluded from the final analysis. The first excluded participant did not perform two cognitive tasks and the chess knowledge task. The second excluded participant did not fill out the survey regarding the estimates for the amount of studying alone and playing chess hours.
The final sample of 77 chess players consisted of 67 males and 10 females, with a mean age of 35 years and an age range of 18 years to 77 years. The chess players had a wide range of chess skills (M = 1683, SD = 574, min = 381, Max = 2651), including several players at grandmaster level. These participants were the same sample as the one used in Lane and Chang's (2018) study.
Procedures
Young Chess Player
Data were collected during one day of informal face-to-face interviews with CS and her parents and one day of testing seven cognitive tasks, one chess memory task, and one chess knowledge task.
The time for CS to complete these cognitive, chess memory, and chess knowledge tasks were the same with others participants. During the interview, CS and her family were encouraged to talk about her experience playing chess, her friendship with other chess players, her hobbies, her entry into playing chess, and her general beliefs about chess.
Information was sought in three important areas. First, we were interested in learning how she had become interested in chess. Second, we wanted to learn more about informal learning experiences that occurred before her first formal engagement with chess activities.
Finally, we wanted to learn about her involvement in any formal or informal training and practice experience after she started playing chess seriously, including her coaching experience, how many hours she had seriously studied alone, and how many hours she had seriously played chess with opponents.
Children
To be eligible to participate in this study, the children had to be 10 years old and have had little or no experience playing chess. The children performed seven cognitive abilities tasks, and it took about 45 to 60 minutes to complete the tasks.
Adult Chess Player
To be eligible to participate in this study, the chess players had to be at least 18 years old and have a USCF rating. Participants were asked to perform the seven cognitive tasks, one chess memory task, one chess knowledge tasks, and fill out a chess-related survey.
The time spent completing the experiment was approximately 120 minutes. Data from all of these tasks were also analyzed in Lane and Chang’s (2018) study.
Materials
In Chang's (2016) dissertation, detailed descriptions and materials for all tasks and surveys were provided. Seven cognitive tasks were used in the study, performed in a specific order: forward digit span, backward digit span, approximate number system, visuo-spatial forward span, visuo-spatial backward span, automated symmetry span, and visual short-term memory task.
1. For the forward and backward digit spans, participants repeated a list of pseudo-random numbers either in the correct order or in reverse immediately after hearing them. The longest length of digits correctly remembered was recorded for each participant for both tasks.
2. The Approximate Number System (ANS) task assessed participants' discrimination sensitivity and internal noise using a computerized task. The Weber fraction (w) was calculated based on individual performances to estimate ANS. A lower w value indicates higher discrimination sensitivity.
3. The visuo-spatial memory spans task uses a computerized version of Corsi's block-tapping task to measure participants' short-term and working memory. Participants recall sequences of colored block locations in the correct order or in reverse, with scores recorded for both tasks.
4. The automated symmetry span task measured participants' dynamic working memory capacity. A higher score indicated higher working memory capacity.
5. The visual short-term memory task evaluated participants' memory using a computerized task with a sequential comparison paradigm. Participants had to detect color changes in arrays of colored squares, with a higher mean d' value indicating a higher visual short-term memory capacity.
Figure 1. Visual short-term memory task.
Chess Knowledge Task
The study used van der Maas and Wagenmakers' (2005) verbal knowledge test, derived from the Amsterdam Chess Test (ACT) and partially adapted by Pfau and Murphy (1988), to examine the chess knowledge of CS and chess players. The test consisted of a total of 18 multiple choice questions of varying difficulty. Participants had to complete this task using pencil and paper with no time limit.
Chess Memory Task
Each position was presented for 10 seconds, followed by a blank screen for 2 seconds. Then the participants were asked to reconstruct the briefly shown positions on an actual chess board. The percentage of accuracy of the reconstructed pieces was calculated.
Chess-Related Background Survey
To ensure continuity with previous research, the simplified version of the survey by Charness et al. (2005) adopted to capture the demographic data of chess players and chess players. This includes (1) demographic information; (2) chess-related development steps; and (3) cumulative and current chess activities.
To understand CS's enjoyment of chess-related activities, their ratings of enjoyment in each of the different chess activities were asked in the questionnaires. She was asked to respond on a 7-point Likert scale, with a rating of 1 being “not pleasant at all” and 7 being “very pleasant.” CS was also asked to estimate her time investment in two categories: (1) serious analysis of chess positions made alone (e.g., with chess books, magazines, databases) and (2) time seriously played with opponents.
The survey questions were asked verbally to CS and the answers were recorded on the questionnaire by the first author. The adult chess players were asked to complete the survey.
Number of games played in Tournaments
CS’s and chess players’ number of games played in FIDE and USCF were obtained from each official website: https://www.fide.com/ and http://www.uschess.org/.
The Variables Associated with Chess
In this study, the measurement of practice in chess-related experience is discussed in comparison to other studies, specifically by Ericsson and Charness. They measured deliberate practice by asking participants to estimate the number of hours spent practicing or studying alone. However, determining what qualifies as deliberate practice can be complex, as studying alone may not always align with Ericsson's definition. This can involve learning chess principles, studying specific opening positions, and preparing moves for known opponents.
For the purposes of this study, the focus is on chess-related experiences such as time spent studying alone, time spent playing non-tournament games, number of tournament games played, and estimated time spent playing tournament games. These experiences encompass deliberate practice and provide an upper limit on the amount of time dedicated to deliberate practice in chess. The study does not delve into the specific definition of deliberate practice, but rather reports results based on these chess-related experiences.
Studying alone
To obtain the total number of study-alone hours, CS and chess players were asked to estimate how many hours they had studied alone for a typical week at a given age. The total number of study-alone hours was calculated by multiplying the weekly estimated average number of hours by the number of weeks in a year and adding the annual estimates at and below that age
Playing non-Tournament Games
In order to determine the total number of hours played seriously with opponents by CS and chess players, participants were asked to provide an estimate of the hours spent playing with opponents in a typical week at various ages. The total number of hours spent studying alone was then calculated by multiplying the average weekly estimate by the number of weeks in a year and adding together the annual estimates for each age or younger.
Playing Tournament Games
CS's and chess players' tournament game data was collected from both the USCF and FIDE official websites. In cases where a game appeared in both databases, it was considered a FIDE game. To standardize the data, the duration of each game was estimated as 4 hours for USCF tournaments and 5 hours for FIDE tournaments. The total number of hours played in USCF and FIDE tournaments was then calculated to determine the overall time spent playing in tournaments.
Chess Experience
Studying alone, playing non-tournament games, and playing tournaments games are all considered chess experience. Thus, the total time for chess experience is the sum of the number of hours of these three chess related activities.
CS’s Chess-Related Background
CS is an only child whose father had a USCF rating of about 1550 when she was 6 years old. At that time, she showed an interest in learning chess and asked her father to teach her the basic rules so she could join the school chess club. Her interest in chess was self-motivated and she started playing seriously at the age of 7. Despite her focus on chess, she still enjoyed and participated in other indoor and outdoor activities.
She received chess instruction both in group settings and individually. At 7 years old, she participated in a chess club and received three months of training during the summer. From age 8 to 10, she occasionally met with a grandmaster for consultation over two to three months. Her chess training was described as relatively unstructured, not very intensive, and irregular. Parts of her interview were documented in Chang's dissertation.
Results
Chess Experience
CS's Chess Rating History shows a clear pattern of rapid improvement. Within 19 months of starting to play chess seriously, she was already at the 88th percentile of the USCF population and continued to climb in rank. After 43 months, she was in the 96th percentile and had defeated an international grandmaster. By 73 months, she was in the 98.8th percentile and had defeated four international grandmasters. This rapid progress calls into question the idea that reaching an international level in chess requires around a decade of experience and intense preparation.
Figure 2. CS’srating history as a function of the number of monthssince she started playing chess seriously. The vertical dashed line indicatesthe interview date. She started playing chessseriously when she
was 7 years and 5 months old, and at the time of this writing, was 13 years and 5 months old. Three types of percentiles are labeled on the right hand-sided y-axis. “All” indicatesthe percentiles were calculated based on the entire population of USCF. “Junior” indicatesthe percentiles were calculated based on all junior
members whose ages were under 21. “U12” indicatesthe percentiles were calculated based on all the child
members whose ages were 12 and below. The smooth curve represents a cubic fit: Rating = 96.12 + 54.34 × Month – 0.868 × Month2 + 0.005 × Month.
CS’s rating history and the percentiles as functions of her amount of chess-related experiences are plotted in Figure 4. Since CS continued to play and study, it is not surprising that the shape of the function relating experience to rating was very similar to that shape of the function relating age to rating: slightly negatively accelerated and close to linear after 1,500 hours of experience.
Figure 3. CS’s estimated accumulated amount of time engaging in chess-related experience by age. “Total” indicates overall chess-related experience.
CS's rate of acquiring chess skills is exceptional compared to other chess players in various studies. For example, compared to players in Gobet and Campitelli's (2007) study who had a mean national rating of 2165 and practiced for about 8,012 hours alone, CS reached a similar level of play after only approximately 3,500 hours of chess-related experience, with a small portion of that time spent practicing alone.
In comparison to players in Charness et al.'s (2005) study who had a mean rating of 2.032 and practiced alone for about 2512 hours, CS reached a comparable level of play after only 156 hours of studying alone, which was 2.4 standard deviations below the mean of Charness et al.'s participants.
Similarly, compared to players in Howard's (2012) study who had a mean rating of 2.122 and practiced alone for about 3981 hours, CS's hours practicing alone were 2.8 standard deviations below the mean of Howard's participants. CS had engaged in approximately 156 hours of studying alone and a total of approximately 3,769 hours of chess experience when she defeated a grandmaster and achieved a high rating placing her in the 96th percentile of the USCF population.
Figure 4. CS’s rating history and percentile as a function of the number of chess-related experience hours. The
dashed circle indicates CS’s rating reached the 96th percentile in the entire USCF population until the interview date,
which was the same day indicated by the vertical dashed line in Figure 2. The smooth curve represents a cubic fit: Rating = 846.6 + 1.2 × Hour – 4.79e-4 × Hour2 + 6.77e-8 × Hour.
In comparison, Dan McLaughlin, who embarked on a journey to become a professional golfer through deliberate practice, reached only around the 90th percentile of the USGA population after approximately 6,000 hours of practice. McLaughlin's performance was far below CS's achievement, and he never reached the level required to play at an international level. Despite engaging in extensive deliberate practice, McLaughlin's performance plateaued at the 90th percentile and did not show the continuous improvement expected based on Ericsson's theory of deliberate practice.
When comparing CS and McLaughlin's improvement rates, it is evident that CS was able to reach high percentiles in chess without extensive deliberate practice, unlike McLaughlin who required thousands of hours of deliberate practice to reach a similar level in golf. The comparison suggests that natural ability may play a significant role in skill development and expertise acquisition. Additionally, the age at which individuals begin their skill development journey, as noted by Gobet and Campitelli (2007), may also influence skill acquisition and development.
Figure 5. CS’s percentiles as a function of hours spent studying alone and chess experience, and McLaughlin’s percentile as a functions of hours of deliberate practice.
This is graphically presented using a smoothing spline with λ, = 0.07.
The data provide two main findings. First, CS's exceptional chess performance calls into question the need for 10 years of dedicated Ericsson training to become an expert. Second, CS shows that she broke the 10,000-hour rule because she was already playing at the international level after gaining less than 4,000 hours of chess experience. This suggests that exceptional performance in chess is not achieved through practice alone. Their performance in chess provides strong evidence that there are individual differences in achieving expertise.
Cognitive Abilities and Chess Knowledge
In the study, CS scored 81% on the chess memory task and 78% on the chess knowledge task, which were much higher than most of the adult chess players in the sample. Only four adult chess players scored higher than CS on the chess knowledge task. Both chess memory and chess knowledge were found to correlate highly with chess rating for the entire sample.
Figure 7 illustrates the accuracy on the chess knowledge test and its relationship with hours spent studying alone and hours of chess experience. Despite studying chess alone for only 156 hours, CS and a few other players were able to obtain high scores on the chess knowledge task. It is noted that very few players with similar study hours and chess experience as CS were able to perform as well in the test. The reason for CS and the few others' exceptional performance with minimal study and experience remains unclear.
Figure 6. The scatterplot of CS and other chess players’ USCF ratings and their accuracy in chess knowledge task. CS’s data is indicated by an empty circle.
Figure 7. The scatterplots of CS and other chess players’ accuracy in chess knowledgeand the numberof hours studying alone and chess-related experience. CS’s data is indicated by an empty circle.
Among the 35 children examined, CS achieved the following ranks in the various cognitive tests: 1st place in visual short-term memory, 3rd place in backward visual-spatial memory span, 4th place in backward digit memory, 6th place in forward visual-spatial memory span , 9th place in the approximate number system, 23rd place in the automated symmetry span and 26th place in the forward digit memory. As shown in the left graph in Figure 8, her score of 2.41 on the visual short-term memory task was significantly higher than the next best child's score.
Descriptive statistics for these children (without CS) and CS on the seven cognitive tests are shown in Table 1.
Table 1. Descriptive Statistics of Seven Cognitive Tasks.
Note: The lower scores in approximate number system task represent better performance
CS performed exceptionally well on a visual short-term memory task compared to 34 other children. To further validate her exceptional performance, an additional 28 10-year-olds were tested using the same procedure.
None of these children, who had not played chess, performed as well as CS. The descriptive statistics for all 62 children showed a mean of 1.24, standard deviation of 0.51, minimum score of 0.08, and maximum score of 2.1.
In comparison to adult chess players, CS's performance was higher than all but three. Overall, CS's performance in the visual short-term memory task was considered exceptional.
Figure 8. Quantile plots of the performance of the three groups in the visual short-term memory task. The "+" sign represents the mean, the "box" extends from the 25th percentile to the 75th percentile, the horizontal line within the box represents the median, the two other horizontal lines represent the inner fences as normally defined in a box plot,
and the horizontal dashed line represents CS’s performance.
It is interesting that both CS and Sammy Reshevsky exhibited excellent performance on visual short-term memory tests, indicating that this ability may be linked to their exceptional chess skills. This led researchers to investigate whether there was a relationship between visual short-term memory and chess skill in adult players. However, a study of 75 adult chess players found no significant correlation between the two variables.
This finding aligns with a previous study by Waters, Gobet, and Leyden in 2002, which also did not find any connections between visual memory and chess proficiency. The researchers did note a small, insignificant negative correlation between experience and visual short-term memory, suggesting that experience may impact the relationship between the two. A multiple regression analysis controlling for chess experience revealed a weak but non-zero relationship between visual short-term memory and chess rating.
This weak relationship leaves open the possibility that CS and Reshevsky's exceptional visual short-term memory may not be directly related to their chess expertise. Researchers also explored potential interactions between visual short-term memory and other variables but found no significant results. It is possible that CS possessed other unmeasured abilities that, when combined with her visual memory, contributed to her chess skill.
The researchers suggest that visual short-term memory may be important for the initial acquisition of chess expertise but may not play a significant role in reaching a high level of proficiency after years of practice. This aligns with Ericsson's argument that factors important for early skill acquisition differ from those important in later stages. The exceptional visual short-term memory of CS and Reshevsky raises questions about the importance of this trait in developing chess expertise.
Further research, especially focusing on young players, is necessary to fully understand the role of visual short-term memory in chess expertise. The findings regarding CS and Reshevsky suggest that this trait may be more significant than previously thought in acquiring chess skills.
| Study 2 |
This study aimed to examine the role of practice in the development of expertise, specifically in the context of chess skills. Chase and Simon's recognition action theory and Ericsson et al.'s theory of deliberate practice were utilized to explore the extent to which practice can explain individual differences in skill acquisition.
The study also investigated the influence of cognitive abilities, such as domain-general fluid intelligence, and chess-specific intelligence on chess skills, considering their contributions after controlling for practice. While deliberate practice theory suggests that practice is the primary factor influencing chess expertise, the study also considered factors such as the amount of time players spent studying alone, playing chess informally, and participating in tournaments.
It was suggested that factors beyond deliberate practice, such as cognitive abilities and chess-specific intelligence, may also play a role in explaining individual differences in chess skills. Ultimately, the study aimed to determine the relative importance of practice and other factors in the development of expertise in chess.
Method
In Study 2, only adult chess players’ data were analyzed. The descriptions of the participants, procedure, and materials of all the tasks and survey were described in Study 1
Measurement
Domain-General Fluid Intelligence.
In this study, the researchers utilized a domain-general fluid intelligence measure that was determined through the performance on seven cognitive tasks. These tasks included tasks such as Digit Span Forward and Backward, Approximate Number System, Block Tapping Forward and Backward, Auto Symmetry Span, and Visual Short-Term Memory.
By analyzing the principal components of these tasks, the researchers were able to establish a measure of domain-general fluid intelligence. This measure provides insight into an individual's cognitive abilities across a range of tasks and can help in understanding overall cognitive functioning.
Chess-Related Fluid and Crystalized Intelligence
Verbal Knowledge Test. The chess-related crystalized intelligence measure was based on performance on fifteen conceptual knowledge questions, and the chessrelated fluid intelligence measure was based on three visualization questions.
Analysis Plan
We planned to conduct a principal components analysis in advance in order to reduce the number of cognitive ability variables and avoid inflating the Type I error rate or reducing power due to multiple tests. The number of components retained was based on the variance explained and ended up being one.
First, we tested for an interaction between cognitive ability and chess experience on chess skill. We then tested the assumption of normality of residuals in the regression model.
A hierarchical multiple regression was used to test the predictions of three views of individual differences in chess skill: deliberate practice, chess experience, and natural ability with previous chess experience. This approach is similar to the one used in Lane and Chang's (2018) study, but in this study, we assessed how much chess knowledge, chess-related fluid intelligence, and domain general fluid intelligence contribute to chess skills after controlling for chess experience.
Result and Discussion
Data were examined carefully to ensure the quality of the data prior to further statistical analyses. The analyses were conducted based on the raw data and on estimates using corrections for unreliability (Cohen & Cohen, 1983).
The corrected estimates were included for two reasons: First, it is not possible to accurately test the magnitude of a correlation between two variables without controlling for the potential distorting effect of measurement error variance second, the analysis based on corrected estimates avoids potential inaccurate estimation in a multiple regression due to the predictors differing in their reliability.
Chess Rating
Chess rating is the dependent variable in the analysis. All participants had a USCF rating, and their current ratings were obtained from the USCF website to avoid potential inaccuracies caused by misremembering or ratings from a different organization. The reliability coefficient used for the chess rating was 91, following Hambrick, Oswald, et al. (2014).
Chess-Related Experience
Both the distributions of the number of hours spent studying alone and playing in non-tournaments were highly skewed, with one player having a value of 0 in both cases. To address this skewness, a log10(x+1) transformation was applied. The reliability coefficient for self-reported studying alone time was .80, based on previous research.
Self-reported playing chess with opponents also had a reliability of 8. The number of tournament games played was determined by adding the number of games played in USCF and FIDE tournaments. A log transformation was used to address the skewness of this variable. Since the number of tournament games played was obtained directly from official chess websites, the reliability of this measure was assumed to be 1.
Cognitive Ability/Domain-General Fluid Intelligence
As outlined in the analysis plan, we predetermined to conduct a principal component analysis to determine if the data from these tasks could be more efficiently analyzed. The results indicated that the first principal component explained 53% of the variance, while the second component only accounted for 13% of the variance. Furthermore, all seven tasks showed strong loadings on the first component, ranging from 0.66 to 0.81 (see Table 2).
These results suggested that a one-component solution effectively summarized the cognitive tasks. This allowed us to carry out further analyses using this primary component instead of individual measures. We named this factor "fluid intelligence" as it encompassed the nature of the tasks, even though it placed more emphasis on memory tasks than typical measures of fluid intelligence.
The domain-general fluid intelligence derived from the principal component analysis of the seven cognitive abilities was deemed highly reliable, with an estimated reliability of 0.9. The second principal component accounted for only 13% of the variance and was not included in subsequent analyses.
Table 2. Loadings of the Seven Cognitive Abilities on the First Principal Component.
Chess-Specific Fluid Intelligence and Crystalized Intelligence.
The reliability of the items for the three chess fluid knowledge questions was 45, and the reliability for the fifteen chess crystallized knowledge questions was 72. Descriptive statistics and a correlation matrix for all variables, including chess rating, seven cognitive abilities, domain-general fluid intelligence, chess-specific memory, chess-specific fluid intelligence, chess-specific crystallized intelligence,
Log time studying alone, Log time playing in non-tournaments, and Log games playing in tournaments, are presented in Table 3.
Two analyses were conducted to evaluate the model. The first tested whether cognitive ability and chess-related experience interacted, but none of the three interactions were significant. The second analysis, which included six predictors in a multiple regression model, did not include the interactions. The residuals of this regression model showed a slight positive skew, but multiple regression is robust even in cases of more positively skewed residuals.
Table 3. The Mean, Standard Deviation, and Correlation Matrix for All Variables
Note. The approximate number system score was rescaled to be positive. The higher scores represent better performance. r’s > .37 are significant with p’s < .001; r’s > .29 are significant with p’s < .01, and r’s > .22 are significant with p’s < .05.
The analyses in the study included both corrected and uncorrected for measurement error using Spearman's disattenuation formula. The corrected and uncorrected correlation coefficients among seven variables are shown in Table 4.
Table 4. Correlations Among Six Tasks and Chess Rating
Note: The reliabilities are shown on the diagonal. The lower triangle contains the correlations corrected for
unreliability and the upper triangle contains the uncorrected correlations.
Hierarchical Regression Analysis
In this study, a multiple regression analysis was conducted to determine if deliberate practice, specifically studying alone, was the sole factor influencing individual differences in achieving chess skills, and to explore other potential factors that may contribute to chess skill. The analysis involved regressing chess rating on six variables, which accounted for 66% of the variance (adjusted R2 = .63) and 82% of the variance (adjusted R2 = .80) after adjusting for unreliability.
This model explained more variance than previous studies, such as Charness et al.'s (2005) model which only accounted for 38% of chess skill variance and Gobet and Campitelli's (2007) model which accounted for 41% of chess skill variance. One possible explanation for this difference is that this study included a wider range of chess ratings than previous studies.
The standard deviations for participants' chess ratings in prior studies varied, with the current study having a standard deviation of 574. Although this was slightly lower than the USCF population's standard deviation of 609, it was still comparable.
Table 5. The Results of the Hierarchical Regression of Chess Rating on the Six Variables
Note: The uncorrected columns are based on the actual data and the corrected columns correct for unreliability. The
column R2 represents the R2 as each variable is entered into the regression model. The column “Min” shows the
proportions of variance explained if the variable is entered last so that none of the common variance is attributed it. The
column “Max” shows the proportions of variance explained if the variable is entered first so that all of the common
variance is attributed to the variable.
A hierarchical regression analysis was used to evaluate the impact of various factors on the prediction of chess skills. The results, presented in Table 5, show the variance explained by each variable individually. The study also corrected correlation coefficients for unreliability.
The analysis revealed that chess experience variables (study alone, informal play, tournament play) accounted for a significant portion of the variance (52% - 60%). Sequentially adding domain-general fluid intelligence, chess-specific fluid intelligence, and chess-specific crystallized intelligence each contributed additional variance to the prediction of chess skills.
The study found that domain-general cognitive ability, chess-specific fluid intelligence, and chess-specific crystallized intelligence all play a role in individual differences in chess skill, even after considering chess experience. While practice is important, deliberate practice alone cannot fully explain differences in chess performance.
The results suggest that deliberate practice, along with other chess experience, is not enough to account for all the variability in chess ratings. Chess knowledge is identified as a crucial predictor of chess skill, indicating that factors beyond practice also play a significant role.
In conclusion, while practice is essential for achieving high levels of expertise in chess, it is not the sole determinant of success. Factors such as cognitive ability and chess-specific intelligence also contribute significantly to individual differences in chess performance.
General Discussion
In 1973, Simon and Chase argued that after 10,000 to 50,000 hours of practice, chess players are able to reach master-level performance by learning thousands of patterns and storing them in long-term memory. Ericsson and his colleagues later emphasized the importance of practice in acquiring a skill, particularly highlighting the concept of deliberate practice, which involves guidance by teachers or coaches.
Deliberate practice is considered the primary source of individual differences in chess skill and has been widely discussed in both popular books and scholarly journals. Studies typically estimate deliberate practice time by asking participants about their weekly hours spent studying chess alone, rather than including time spent playing chess.
The study presented in this text highlights a chess player called CS, who achieved impressive results after only 3 years and 7 months of playing chess seriously and spending approximately 156 hours studying chess. However, it remains unclear if CS met Ericsson's criteria for deliberate practice, as defined in previous research.
Study 2 further illustrates that chess experience, including time studying alone, is necessary but not sufficient for developing expertise in chess. Other factors such as domain-general fluid intelligence, domain-specific fluid intelligence, and chess-specific crystallized intelligence also play a role in chess skill development.
The findings suggest that chess expertise is not solely dependent on practice, challenging the 10-year rule or 10,000-hour rule. While practice remains essential, other factors such as deep search ability, strategic planning, and positional principles also contribute to a player's skill level.
Overall, the results support the idea that practice is necessary but not sufficient for achieving high levels of expertise in chess. Further research is needed to explore the additional factors that influence chess skill development beyond deliberate practice and chess experience.
| Conclusion |
The theory of deliberate practice proposed by Ericsson and the recognition-action theory proposed by Chase and Simon both suggest that the key to success lies in deliberate practice. Reaching master-level performance in chess requires at least 10 years or 10,000 hours of deliberate effort, according to Ericsson. The primary source of individual differences in chess skill is deliberate practice.
Study 1 is focusing on a talented young chess player, CS, it was found that despite spending relatively little time studying alone, she achieved remarkable proficiency in chess. This challenges the idea that 10 years or 10,000 hours of practice are necessary to excel in chess. CS's exceptional performance in chess, compared to other players in other studies, suggests that natural ability may play a significant role in skill development.
Study 2 further explored the role of practice in developing expertise in chess. The findings indicated that while practice is important, factors such as cognitive abilities and chess-specific intelligence also contribute significantly to individual differences in chess skill. Despite the importance of deliberate practice, it is not the sole determinant of success in chess.
Overall, the research suggests that while practice is crucial for skill acquisition in chess, other factors such as cognitive abilities, chess-specific intelligence, and natural ability also play a significant role in achieving expertise. The results challenge the notion that practice alone is sufficient for reaching high levels of proficiency in chess and highlight the complex interplay of various factors in skill development.
