Introduction
Chess is often considered a “drosophila” of cognitive psychology, given the extensive interest in understanding the basis of expert performance levels in this field. For decades, chess research has demonstrated that high-level performance is strongly supported by a large repository of domain-specific knowledge acquired through deliberate and structured practice. However, the extent to which general individual characteristics, especially intelligence, contribute to the acquisition and maintenance of this expertise remains controversial. While some authors argue that intelligence quotient (IQ) is irrelevant or only weakly related to expert performance (Ericsson & Lehmann, 1996), other studies suggest that intelligence can explain a significant portion of the variance in high-level achievement (Ceci & Liker, 1986; Brody, 1992). In this context, Grabner, Stern, and Neubauer (2007) conducted a comprehensive psychometric study to investigate the relationship between several components of intelligence, personality variables, chess experience, and the level of expertise in adult tournament players.
Objectives
The authors established three main objectives:
1. To evaluate the association between different components of intelligence (verbal, numerical and figural) and the level of chess expertise, measured by the ELO ranking of the participants.
2. To examine the contribution of chess experience—both in terms of current practice (including deliberate practice activities) and prior trajectory (age at start, length of club membership, number of tournament games)—to chess performance.
3. To explore the influence of personality traits and emotional competencies (through standardized questionnaires) in predicting the level of expertise.
Method
Participants: 90 tournament chess players (87 men and 3 women), aged 15–65 years (M = 36.23; SD = 13.29). ELO rankings ranged from 1311 to 2387 (M = 1869; SD = 247), ranging from intermediate to potential grandmasters.
Intelligence instruments: The Intelligenz-Struktur-Test 2000 R (IST 2000 R; Amthauer et al., 2001) was used, which provides standardized scores (M = 100; SD = 15) in the following domains:
Verbal Intelligence: linguistic functioning (sentence comprehension, analogies and verbal similarities).
Numerical Intelligence: performance in arithmetic tasks, number series and mathematical operations.
Figural Intelligence: visuospatial abilities (figure selection, mental rotation of cubes and matrices).
General Intelligence (g): index composed of the three domains above.
Chess questionnaire:* Included sections on biographical data (age at first playing chess, club membership, active breaks), tournament game record (number of games and percentage of results in 2003-2004), attitude toward chess (subjective importance, enjoyment of playing, motivation for practice and performance), and estimated hours per week spent on different practice activities (individual practice with books, use of software, practice with partners, informal games, lessons given, lessons received, coverage of tournaments in the media). Each activity was assessed in terms of *deliberate practice* following the criteria of Ericsson et al. (1993): degree of contribution to performance improvement, effort required, and inherent enjoyment.
Personality and emotional competencies:
NEO-FFI Personality Inventory (Costa & McCrae, 1989; translated by Borkenau & Ostendorf, 1993): assesses five factors (neuroticism, extraversion, openness, agreeableness and conscientiousness), with standardized T-scores (M = 50; SD = 10).
Emotional Competence Questionnaire (FEK; Freudenthaler & Neubauer, 2005): measures self-assessments in six areas: perception of one’s own emotions, perception of others’ emotions, control of emotional expression, masking of emotions, regulation of one’s own emotions and regulation of others’ emotions, also in T-scores (M = 50; SD = 10).
Procedure: Tests applied to groups of 2 to 14 participants, in a controlled environment, with no time limit for questionnaires. The complete battery took approximately 3 hours.
Data analysis: Pearson correlations (or Spearman correlations when necessary) were computed between ELO and each variable, in addition to multiple linear regressions to verify independent contributions. The significance level was set at 5%.
Results
1. Intelligence profiles
Means compared to normative (IST 2000 R):
General Intelligence (M = 113.53; SD = 14.05) was significantly above the normative mean (100), t(89) = 10.60, p < 0.01.
Numerical Intelligence (M = 116.41; SD = 14.15) showed the greatest increase, approximately one standard deviation above expected, t(89) = 12.02, p < 0.01.
Verbal Intelligence (M = 108.41; SD = 13.36) and Figural Intelligence (M = 106.14; SD = 15.41) also exceeded the normative ones, both p < 0.01.
Correlation between IQ and ELO:
Numerical Intelligence: r = 0.46, p < 0.01 (highest correlation);
Verbal Intelligence: r = 0.38, p < 0.01;
General Intelligence (g): r = 0.35, p < 0.01;
Figural Intelligence: r = 0.02, not significant.
Subscales: Among the subscales, the “Number Series” showed the highest correlation with ELO (r = 0.44, p < 0.01), suggesting that inductive reasoning in the numerical domain is moderately well related to playing strength. The figure selection and mental rotation of cubes subscales (purely visuospatial components) did not show significant correlations with ELO, indicating that classical spatial manipulation skills do not directly predict the level of chess expertise.
2. Chess experience and tournament activity
Biographical data:
Age at first chess entry: r = –0.38, p < 0.01 (the earlier, the higher ELO).
Age at club membership: r = –0.50, p < 0.01 (earlier entry into clubs associated with better level).
Recent tournament activity: Number of matches played in 2003-2004: r = 0.45, p < 0.01 (number of matches positively correlates with ranking).
Percentage of positive results: r = 0.50, p < 0.01 (better performance is reflected in higher ranking).
Independent contributions: Although a strong relationship between number of matches and performance was expected, the partial correlation between ELO and number of matches, controlling for tournament effectiveness, remained significant (r = 0.34, p < 0.01), suggesting that simply playing more matches in tournaments is associated with expertise, independent of specific results.
3. Attitudes and motivation towards chess
Subjective importance of chess (how much the individual values the game in his or her life): r = 0.28, p < 0.01;
Chess performance motivation (desire to outperform others and self-demand): r = 0.39, p < 0.01.
Enjoyment of playing and motivation for deliberate practice* were not significantly correlated with ELO.
Although 66% of participants reported practicing with written material for an average of 1.5 hours/week and 57% played with others, none of the estimates of actual practice hours (individual or group) correlated positively with ELO. Instead, such estimates were more closely associated with level of motivation to practice (r = 0.43, p < 0.01), suggesting that self-assessments of practice may reflect attitude rather than actual practice.
4. NEO-FFI (Big Five) personality traits and emotional competencies
: None of the Big Five factors (neuroticism, extraversion, openness, agreeableness, conscientiousness) correlated significantly with ELO.
FEK (Emotional Skills): The only subscale that was significantly related was control of emotional expression (r = 0.27, p = 0.01), indicating that stronger players tend to have greater ability to control what they demonstrate emotionally during the game.
5. Multiple regression
Variables that entered the regression model to predict ELO (adjusted R² = 0.55; F\[6,81] = 18.65, p < 0.01):
1. Age at club affiliation (β = –0.44; p < 0.01);
2. Number of matches in tournament (β = 0.33; p < 0.01);
3. Control of emotional expression (FEK) (β = 0.20; p < 0.01);
4. Numerical Intelligence (β = 0.31; p < 0.01);
5. Current age (β = 0.28; p < 0.01);
6. Motivation to perform well at chess (β = 0.17; p < 0.05).
In another model restricted to intelligence measures only, Numerical Intelligence was the only significant predictor, explaining about 22% of the variance in ELO (adjusted R² = 0.20; F\[1,88] = 12.15, p < 0.01).
Discussion and Implications
1. Do chess players have high IQs?
In absolute terms, the sample of tournament players presented means for general, verbal, and numerical IQs above the normative mean. In particular, the numerical component showed a standard deviation of approximately +1 SD, indicating notable superiority in arithmetic and numerical reasoning tasks compared to the general population.
The moderate correlation between General IQ (r = 0.35) and, more importantly, Numerical IQ (r = 0.46) and ELO ranking suggests that stronger tournament chess players tend to have above-average general intellectual abilities, especially in the numerical domain. Although the magnitude of this correlation is not very high, it is consistent and statistically significant, indicating that chess requires a certain level of general cognitive ability, and does not operate in isolation from other higher mental abilities.
In contrast, the conventional visuospatial component (Figure IQ), measured by mental rotation and figure selection tests, did not correlate with ELO. This implies that, although visuospatial processes are fundamental to understanding and manipulating patterns on the board, the skills assessed by traditional visual-spatial IQ tests do not capture the specific competencies involved in chess, which probably depend on knowledge structures of piece positions and pattern recognition, rather than pure mental rotation.
2. Integration between experience and intelligence
The strong contribution of age at entry into a club and number of tournament matches reinforces the idea that long-term experience is crucial to achieving high levels of performance (Ericsson & Lehmann, 1996). However, the fact that Numerical IQ still explains a significant portion of the variation in ELO, even after controlling for experience variables, indicates that more intelligent players tend to benefit more, in terms of converting experience into superior performance.
This combination of experience and cognitive skills suggests that, although chess is a domain that can be learned through deliberate practice, there is a certain cognitive threshold that facilitates the development of expertise. However, players who have surpassed 2200 ELO have numerical IQs of around 110–115, demonstrating that it is not necessary to have an extremely high IQ to become an expert; it is enough to have a *slightly above average IQ* and to invest in consistent practice.
3. Role of emotional skills
The correlation between control of emotional expression and ELO (r = 0.27) indicates that stronger players have a greater ability to externally manage their emotions, which can help them maintain calm under pressure and make more rational decisions during critical matches. This ability is not necessarily synonymous with “being cool”, but rather with strategic emotional regulation, preventing spontaneous reactions from negatively influencing reasoning.
4. Practical implications
For coaches and chess clubs, the results point to the importance of identifying young players with good cognitive potential, especially in numerical reasoning, and at the same time promoting guided deliberate practice (use of teaching materials, analysis of high-quality games and systematic training) from an early age.
Developing training programs that strengthen emotional control during play (through tournament simulations, learning breathing techniques, and reflecting on adverse scenarios) can contribute to better performance in high-level tournaments.
5. Limitations and future perspectives
Since this is a cross-sectional study, it is not possible to determine the causal direction of the relationship between IQ and chess performance. Longitudinal studies that follow the evolution of children and adolescents in chess, correlating IQ trajectories and changes in ELO over time, would be essential to clarify whether high IQ facilitates progress or whether intense chess practice develops certain specific cognitive skills.
Further exploring the visuospatial aspects of chess (e.g., piece patterns, memorization of tactical structures) through cognitive tests more closely tailored to this domain could shed light on why traditional figural IQ did not correlate with ELO. Perhaps the assessment tool needs to capture components such as positional memory and strategic pattern recognition, not just mental rotation.
Future samples should include a larger number of grandmasters and examine the possible existence of IQ plateaus at very high ELO levels, as well as investigate the interaction between genetic factors (such as variations in genes related to cognitive processing) and environmental variables (coaches, quality of clubs, access to resources).
Conclusion
The study by Grabner, Stern and Neubauer (2007) robustly demonstrates that tournament chess players have, on average, a general IQ and, especially, a numerical IQ* above the population average, and that these general cognitive abilities are moderately associated with the level of expertise, as measured by the ELO ranking. Furthermore, it indicates that long-term experience, regular practice in tournaments and emotional skills (such as control of emotional expression) are also relevant determinants. Therefore, chess players do not have “very high” IQs in absolute terms (i.e., they are not generally gifted in the strict sense), but they tend to have slightly above-average IQs, which, combined with a consistent trajectory of practice and effective emotional regulation, contribute to achieving high levels of performance.
Personal observation: I noticed, when reviewing studies correlating cognitive abilities and performance in specialized domains, that traditional IQ measures often do not capture specific nuances of each area, as is the case with visuospatial ability in chess.
Reference
Grabner, R. H., Stern, E., & Neubauer, A. C. (2007). Individual differences in chess expertise: A psychometric investigation. Acta Psychologica, 124(3), 398–420. doi:10.1016/j.actpsy.2006.07.008
