Acceptance and Commitment Therapy for Improving the Performance of Chess Players Suffering from Anxiety Disorders

Acceptance and Commitment Therapy for Improving the Performance of Chess Players Suffering from Anxiety Disorders

Full reference: Ruiz, F. J., Luciano, C., & Suárez-Falcón, J. C. (2023). Acceptance and Commitment Therapy for Improving the Performance of Chess Players Suffering from Anxiety Disorders. International Journal of Psychology & Psychological Therapy, 23(2), 207-219.

Study type: Multiple case study with matched control participants.

Background and objectives

Previous research has suggested that brief protocols based on Acceptance and Commitment Therapy (ACT) are efficacious in improving elite chess players' performance without clinical problems. The authors argue that these promising results warrant examining the efficacy of longer ACT interventions with elite chess players suffering from emotional difficulties. This study advances in this direction by presenting two case studies of elite chess players suffering from high anxiety levels, comparing their evolution with matched control participants. The specific objective was to present the effect of intensive ACT interventions (five 90-minute sessions) on chess performance of two elite players suffering from high levels of anxiety.

ACT is a contextual-behavioral approach to psychological intervention that promotes psychological flexibility—the ability to stay in the present moment, mindfully aware of private events (thoughts, memories, sensations, emotions), and committed to valued goals. The authors recognize the role of psychological inflexibility and experiential avoidance in chess competitions, where decision-making is sometimes controlled by cognitive fusion and experiential avoidance processes that deteriorate performance.

Method

Participants

Case 1 (Raúl): Participant 1 was Raúl, a 23-year-old chess player considered one of the most talented players in Spain. His progress had been blocked for the last years before the intervention. Raúl received psychological treatment based on rational emotive behavior therapy for obsessive-compulsive disorder (OCD) and social anxiety disorder during the year immediately before the beginning of the intervention. However, he did not experience significant improvement.

Case 2 (Mario): Participant 2 was Mario, a 12-year-old boy with high performance anxiety during chess competitions. He was categorized as an exceptionally gifted boy and had played chess since early childhood, obtaining excellent results at a national level. However, he experienced a significant decrease in his performance after being told by a renowned Spanish chess player that he was the most talented young player. Mario and his parents consulted with a renowned chess trainer for his anxiety problem, but the guidelines provided did not show the desired effect.

Matched control participants: Control participants were matched with each participant by searching the International Chess Federation (FIDE) database. The criteria to find control participants were as follows: (a) same gender, (b) age difference of no more than one year, (c) same International Title, (d) difference in ELO score of no more than 30 points, (e) equivalent frequency in playing competitions during the last 12 months, and (f) playing in the same geographic zone (e.g., tournaments in the south of Spain). Only one chess player was matched to every participant because the precise specification of the criteria made it difficult to match both experimental participants to more than one person.

Design

The study employed a multiple case design with matched control participants. Chess performance of all participants was monitored during an 18-month follow-up period. It was expected that chess players would show a relatively stable performance around a mean baseline level (i.e., there would not be significant baseline trends) and that the intervention would produce an immediate and stable change in chess performance.

Intervention / Conditions

Case 1 (Raúl): The intervention with Raúl consisted of five 90-minute sessions conducted on the same number of consecutive days plus brief periodic contact (approximately one per two months) for the next year after the intervention.

Session 1 was dedicated to conducting a functional analysis of the problem and promoting creative hopelessness. The aim was for Raúl to experience the pernicious consequences of trying to control and avoid obsessive thoughts and social anxiety. In reaction to these problems, Raúl was invited to behave fused with many persecutory thoughts in his daily life and chess competitions. In reaction to those, Raúl was invited to engage in rituals. He also fused his behavior with social anxiety by leaving the situation when it felt like it was most vital for him to be there.

Sessions 3 to 5 were dedicated to providing multiple-exemplar training in defusing from obsessive thoughts and anxiety while engaging in committed actions. These included "take your mind for a walk" exercises (Hayes et al., 1999, p. 161; Wilson & Luciano, 2002, p. 199), "taking your mind for a walk exercise" (Hayes et al., 1999, p. 162; Wilson & Luciano, 2002, p. 215), engaging in social interactions on campus that involved anxiety (e.g., asking for store information, making infrequent questions, playing the chessboard metaphor game (Hayes et al., 1999, p. 191; Wilson & Luciano, 2002, p. 209). Raúl was also invited to imagine his own funeral and, first, think about who would come and what they would say about how his life had been until that moment and, second, what he would like them to say.

Case 2 (Mario): The intervention with Mario was also programmed to be implemented intensively in five 90-minute sessions conducted within three days, with new contacts to review his progress every two months, approximately, during the next year.

An initial assessment session was conducted with Mario and his parents. Mario presented multiple problems during chess competitions related to psychological inflexibility: (a) high anticipatory anxiety that usually results in a 30-45 minute deficit after the beginning of each game; (b) he usually gets into trouble because he was afraid of making errors (e.g., seeing his mother in the audience); (c) when he had anxiety, he often focused on physical details of anxiety rather than details specifically related to the task (e.g., you have many options); (d) he avoided moves that lead to tactical complications because he had a clear advantage; (e) he avoided moves that lead to tactical complications because he had a clear advantage; (f) he often offered draws in favorable positions; (g) he played defensively after losing a game to avoid losing another game; (h) he used self-talk too much to calm himself (e.g., thinking that he was playing a friendly game); (i) when he lost a game, he wanted to quit chess, blamed his mother, and committed to himself not to do this; and (j) before games, he checked whether all these thoughts surfaced and whether he was having good moves.

Session 1 was dedicated to promoting creative hopelessness by making a diagram with Mario. Then, creative hopelessness was promoted by suggesting that Mario's behavior pattern be compared to a turtle whose life mission was to walk north. However, when some problems appeared (e.g., rain, insects, etc.), he stopped walking and tried not to get into his shell. Raúl strongly identified with this metaphor. Finally, to clarify values, the "funeral exercise" was introduced in which Mario was invited to imagine his own funeral and, first, think about who would attend and what they would say according to what his life had been like up to that moment and, second, what he would like them to say.

During the rest of the intensive intervention described above, several sessions were conducted with Mario and one of his parents to review Raúl's advantages, present relapse, and explicitly connect the work conducted promoting psychological flexibility in daily life with chess competitions. In a telephone interview maintained with Raúl at the beginning of his behavior in other areas became more flexible, and he stated, "I have stopped behaving artificially, and I feel more natural now." The change developed during the first months after the intervention in which the following guidelines of his "mind" did not generate him great confusion and distress and, despite them, he behaved according to his values.

After the intensive intervention described above, several sessions were conducted with Mario and one of his parents to review the advantages, present relapse, and explicitly connect the work conducted promoting psychological flexibility in daily life with chess competitions. In a telephone interview maintained with Mario at the beginning of his behavior in other areas became more flexible. Immediately after the intensive intervention, Mario experienced an improvement in chess performance. However, a relapse occurred approximately after eight months. It is worth noting that this deterioration in chess performance occurred in tournaments that were especially difficult for the player and that the other therapeutic games were maintained and even increased during this period. After additional sessions, the intervention was considered finished, and Mario showed a significant increase in chess performance during the 18-month follow-up.

Outcome measures

ELO Performance Rating (Elo, 1978): ELO Performance scores were used as the primary outcome measure. ELO Performance is a theoretical index primarily used by FIDE to establish chess skill levels through consecutive games. ELO Performance scores were obtained for each participant and their matched control pairs by analyzing data from competitions provided by FIDE's database, in which all participants played.

Acceptance and Action Questionnaire-II (AAQ-II): Bond et al., 2011; Raith, Langer, Luciano, Cangas & Beltrán, 2013. The AAQ-II is a general measure of psychological inflexibility. The AAQ-II is a 7-item Likert scale with 7 points. Higher scores indicate higher levels of psychological inflexibility. The AAQ-II was used only with Participant 1.

Avoidance and Fusion Questionnaire-Youth (AFQ-Y): Greco, Lambert, & Baer, 2008; Spanish version by Salazar et al., 2019. The AFQ-Y is a 17-item scale, 5-point Likert scale. It measures psychological inflexibility in children and adolescents. Higher scores indicate higher levels of psychological inflexibility. The AFQ-Y was only administered to Participant 2.

Kentucky Inventory of Mindfulness Skills (KIMS): Baer, Gregory, & Allen, 2004. The KIMS is a 39-item scale that measures four mindfulness skills: Observing, Describing, Acting with Awareness, and Accepting without Judgment. Higher scores indicate greater mindfulness skills. The Spanish translation by Ruiz (2014) was used, which showed good internal consistency.

Chess Counterproductive Reactions Questionnaire (CCRQ): Ruiz & Luciano, 2009, 2012. The CCRQ is a 15-item, 9-point Likert scale designed to detect psychological inflexibility in chess competitions. Preliminary data show it has good internal consistency and correlates with AAQ-II scores and ELO ratings (Suárez-Falcón, Ruiz, & Luciano, 2012). The CCRQ was applied only to Participant 1.

Psychological Barriers in Chess Questionnaire (PBCQ): The PBCQ is a 30-item, 5-point Likert scale. It measures competitions and the flexible reaction to them. It was designed based on the CCRQ. Preliminary data show that the PBCQ has adequate internal consistency (Suárez-Falcón et al., 2012). The PBCQ was administered only to Participant 2.

Self-monitoring: Participant 1 was invited to complete a self-registry after each competition game. Specifically, he responded to what degree he experienced perfectionist thoughts (0-10) and whether he had controlled his play. Participant 2 was invited to fill out a similar register, but because there was no possibility of obtaining a baseline in this case, these data are not presented here.

Data analysis

Two methods of analysis were used that were considered to adequately fit these previous assumptions: (a) JZS+AR Bayesian hypothesis testing for single-subject designs (de Vries & Morey, 2013) and (b) nonparametric Tau-U statistic (Parker, Vannest, & Davies, 2011). The JZS+AR Bayesian model is useful when data within each phase are expected to be stable around a certain true mean (i.e., there is no trend), and the intervention is assumed to produce a level change in scores right after intervention implementation. This model adapted the JZS t-test and accounts for the serial dependence typical of single-subject designs with an autoregressive [AR(1)] model. The JZS+AR model provides a Bayes factor (BF) that quantifies the relative evidence in the data for the hypothesis of no intervention effect and for the hypothesis of intervention effect (i.e., true effect). Additionally, this model estimates effect size by standardizing the difference in true means between phases. This standardized mean difference coefficient, termed δ, is slightly different from conventional Cohen's d, where the mean difference is standardized by the within-group standard deviation.

To analyze the appropriateness of the JZS+AR model, Tau-U tests were first computed for baseline and follow-up to establish whether there were significant within-phase trends in participants' performance. Subsequently, a Tau-U value was computed for the effect of intervention on each participant. Tau-U values were calculated using the online calculator provided by Vannest, Parker, and Gonen (2011).

Results

Case 1 (Raúl)

Figure 1 shows Raúl's evolution in ELO Performance on the left and his matched control participant on the right. Raúl did not show a significant tendency across baseline (Tau-U = -.039, Z = -.374, p = .708), nor during the 18-month follow-up (Tau-U = -.227, p = .802). However, according to the JZS+AR Bayesian model, a level change in ELO Performance occurred after the intervention. Specifically, Raúl performed at a mean level of 2437.3 points (SD = 89.543) during baseline and at a mean level of 2527 points during the 18-month follow-up (M = 2527, SD = 60.672, δ = 1.064, ELO Performance increase of 89.7 points). The BF was 53.5, which means that the data support the hypothesis of intervention effect by this ratio compared to the hypothesis of no intervention effect. The nonparametric Tau-U statistic also indicated an intervention effect regarding ELO Performance (Tau-U = .606, SEtau = .163, Z = 3.712, p = .0002).

Raúl's matched control participant showed a small but nonsignificant positive trend during baseline (Tau-U = .21, Z = 2.013, p = .04) but not during the 18-month follow-up (Tau-U = .033, Z = .18, p = .857). Although the positive trend during baseline was significant, we decided not to control for the trend given the small value of Tau-U, as suggested by Parker et al. (2011), and continue applying the JZS+AR Bayesian model. Concerning this model, control participant 1 performed at a mean level of 2420.2 ELO Performance points (SD = 94.558) and at 2414.9 points at the 18-month follow-up (SD = 93.197, δ = -.035, ELO Performance decrease of 5.3 points). The BF was 0.234, which means that the data support the hypothesis of no improvement by a factor of 4.258 compared to the hypothesis of improvement. The Tau-U also indicated no improvement between baseline and follow-up (Tau-U = -.054, SEtau = .17, Z = -.318, p = .751).

Regarding perfectionist thoughts frequency, a significant frequency of perfectionist thoughts was not found during the baseline (Tau-U = -.218, Z = -.934, p = .35), nor during the follow-up (Tau-U = .044, Z = .51, p = .61). A significant intervention effect was identified (Tau-U = .219, Z = 2.019, p = .0249). Regarding flexible reactions, there were no significant tendencies during the baseline (Tau-U = .036, Z = .156, p = .876) and during the follow-up (Tau-U = .071, Z = .562, p = .574). According to Tau-U, the intervention increased the frequency of flexible reactions in response to perfectionistic thoughts (Tau-U = .493, Z = 2.598, p = .009).

Self-report measures for Case 1: Table 1 shows Raúl's scores on the self-report measures at pretreatment and the 18-month follow-up. Psychological inflexibility levels decreased from a clinical to a nonclinical score. Likewise, chess-related experiential avoidance significantly decreased. All mindfulness measures increased, especially the score on the KIMS factor of Accept without Judgment.

Case 2 (Mario)

Figure 3 presents the evolution in ELO Performance of both Mario on the left and his matched control participant on the right. Mario did not show a significant tendency across baseline and treatment period (Tau-U = -.078, Z = .508, p = .612), nor during the 1-year follow-up (Tau-U = .111, Z = .447, p = .655). According to the JZS+AR Bayesian model, Mario performed at a mean level of 1838.3 ELO Performance points (SD = 173.776) during the baseline and treatment period and at a mean level of 2006 ELO Performance points during the 15-month follow-up (SD = 441.171, δ = 1.104, ELO Performance increase of 167.7 points). The BF was 6.2, which means that the data support the hypothesis of intervention effect by this factor compared to the hypothesis of no intervention effect. Likewise, the Tau-U statistic indicated an intervention effect regarding chess performance (Tau-U = .673, SEtau = .224, Z = 3.009, p = .003).

Mario's matched control participant showed small but nonsignificant positive trends during baseline (Tau-U = .225, Z = 1.216, p = .224) and the 1-year follow-up (Tau-U = .25, Z = .866, p = .387). According to the JZS+AR Bayesian model, control participant 2 performed at a mean level of 1977 ELO Performance points (SD = 106.237) and 2035.9 points at the 1-year follow-up (SD = 128.534, δ = 0.491, ELO Performance increase of 58.9 points). The BF was 0.637, which means that the data support the hypothesis of no improvement between baseline and follow-up by a factor of 1.569 compared to the hypothesis of improvement. The Tau-U also indicated no improvement between baseline and follow-up (Tau-U = .477, SEtau = .255, Z = -1.868, p = .062).

Self-report measures for Case 2: Table 1 shows Mario's scores on the self-report measures at pretreatment and the 1-year follow-up. According to the PBCQ, the frequency of problematic private experiences when playing chess significantly decreased from pretreatment to posttreatment and follow-up, while flexible reactions increased. In addition, psychological inflexibility levels decreased from a clinical to a nonclinical score. Mario also showed clinically significant increases in Act with Awareness and Accept without Judgment regarding mindfulness skills.

Discussion and conclusions

Chess players' decision-making is sometimes subjected to the influence of cognitive fusion and experiential avoidance, which usually deteriorates their performance. The identification of these processes permitted designing brief ACT protocols that proved to be efficacious in improving elite chess players' performance without clinical problems (Ruiz, 2006; Ruiz & Luciano, 2009, 2012). These promising results of brief ACT protocols warrant examining the efficacy of longer ACT interventions with elite chess players suffering from emotional difficulties. Accordingly, the current study aimed to present the results obtained by applying ACT to two elite chess players suffering from high anxiety levels.

Both interventions resulted in treated participants experiencing clinically significant reductions in symptomatology and improved valued living during follow-up after the intervention. Both treated participants significantly improved their chess performance compared to the two control participants. The control participants did not significantly improve their chess performance during follow-up, but the treated players experienced substantial improvements in their chess performance.

An important limitation and strength of the current study merits mentioning. First, the main limitation is that only two chess players received the intervention. However, on the other hand, it is important to consider that the potential number of participants is very low because, by definition, there are few chess players, and only some have experience with psychological problems. On the other hand, single-case experimental designs seem to be the most appropriate methodology for analyzing the effect of these processes. Second, the matched control participants had no contact with the experimenters. Accordingly, we cannot ensure that only possible positive expectations and greater motivation would lead participants to improve their performance. To balance these limitations, this study was conducted considering Martín, Valiente, and Schwartzman (2005) recommendations for research in sport psychology. In this case, participants were chess players who competed on a regular and organized basis, performance was measured directly with a highly reliable and valid chess performance measure, and the follow-up was unusually long.

In conclusion, this study provides additional evidence of the potential of ACT to improve elite chess players' performance, in this case with those experiencing high anxiety levels. Future research might conduct a more systematic evaluation of the effect of ACT interventions in increasing chess players' performance suffering from psychological disorders.

Significance and Contribution

This study provides empirical evidence that intensive ACT interventions can improve chess performance in elite players experiencing anxiety disorders. Unlike previous research limited to brief interventions with non-clinical players, this multiple-case study with matched controls demonstrates that extended interventions (five 90-minute sessions plus periodic follow-up) effectively address both psychological symptoms and sport performance in clinically compromised elite athletes. Both treated participants showed clinically significant reductions in emotional symptoms and improvements in valued living, accompanied by substantial chess performance gains that were not replicated in matched controls. The study addresses a gap in sport psychology by integrating ACT with competitive sport contexts, demonstrating that psychological interventions focused on increasing psychological flexibility can enhance elite performance in clinically affected athletes.

---

---