Development of Specific Motor Skills through System Wall Bouldering Training: A Pilot Study

This study evaluated the effects of a five-week period of practicing specific climbing movements using a system wall on motor skills and bouldering performance compared to self-regulated, conventional bouldering. Thirteen advanced female boulderers (age: 24.5 ± 3.6 years, height: 166.9 ± 3.4 cm, and body mass: 63.4 ± 8.0 kg) were divided into an experimental group (n = 7) and a control group (n = 6). Both groups continued their normal training routines during the intervention, but the experimental group dedicated 30 minutes of their climbing time twice per week to practicing specific motor skills on a system climbing wall. Before and after the intervention, the participants attempted two boulder problems on the same wall. The performance was registered as the number of attempts to complete the boulder problems and as the highest hold reached within four attempts. Video recordings of climbers' best attempts, capturing the highest hold reached from a perspective directly behind them, were analyzed by three independent experts. The analysis was conducted using a five-point scale across six categories of movement quality. Modest enhancements in certain motor skills and performance were evident in both groups, revealing no significant distinction between them. The results underscore the efficacy of incorporating system walls into the training routines of advanced female boulder climbers, but the absence of between-group differences highlights the significance of individual preferences when choosing between conventional and system wall bouldering.


Introduction
Recent years have witnessed an increased interest in climbing among professional and recreational circles.While the enhancement of sport-specifc ftness in male climbers at higher performance levels has been extensively examined [1], there is a notable gap in research concerning recreational female climbers [2].Tis group's engagement with resistance training is lower [3,4], underscoring the need to explore climbing-specifc activities that could improve their performance.Te importance of strength is recognized [5][6][7], yet the critical role of motor skills, for instance, executing efcient movements and adjusting body positions swiftly [8,9], should not be overlooked [10].
In modern commercial bouldering gyms, the volume of available problems provides climbers with a wide array of challenges.While this variety may encourage exploration of diferent climbing styles and movement types, it can also inadvertently steer climbers to seek routes that align with their preferred climbing variety (e.g., hold types and wall steepness).Tis natural inclination towards familiar types of climbing, coupled with the vast selection, may lead climbers to engage less frequently in the repetitive practice of specifc problems in their self-regulated training.Such repetition is essential for errorless learning and motor skill refnement [11][12][13].Although climbers probably revisit problems they have completed, either for enjoyment or as part of their training regimen, the tendency to continuously seek new challenges might limit the opportunities for focused skill development through repeated attempts at the same problem.Previous studies suggest that structured climbing training can enhance performance without necessarily increasing the frequency or training intensity [14,15].Leveraging the diversity of problems while also encouraging targeted repetition could address the gaps in current practice and promote more comprehensive skill development among climbers.
Recently, walls equipped with lights integrated in the holds and adjustable steepness have emerged as novel training tools in bouldering.Such walls are capable of displaying a multitude of community-created boulder problems with consensus-based difculty rating which can be accessed by a phone application.Tis type of wall presents a potentially useful tool for standardizing performance measures, quantifying and tailoring training, and adapting the difculty and style to all levels of climbers.Using technologically augmented walls for climbing training may hold potential, especially if their efectiveness in enhancing motor learning and skill development can be demonstrated.By incorporating automated, visual instruction in climbing training, it further becomes possible to optimize the allocation of human resources by reducing the hands-on time required by a trainer.Despite their potential benefts, such walls remain underexplored in research.
Acknowledging the gap in climbing literature on movement skill development and recreational, female climbers [2,8], our pilot study evaluates the efects and feasibility of a fve-week training period using a system wall on motor skills and bouldering performance compared to the same duration of self-regulated, conventional bouldering training.Given the theory that structured, repetitive practice can lead to signifcant improvements in climbing-related ftness and performance [11,12,14,15], we hypothesized that the deliberate practice on prescribed boulder problems on a system wall would result in greater improvements in motor skills and bouldering performance compared to the self-regulated bouldering training.

Experimental
Design.An experimental trial with two groups was employed to investigate the research question.Prior to the commencement of the intervention, all participants were assessed for bouldering performance and technical execution.Te group assignment was determined by randomly drawing lots from an opaque container immediately following the completion of all pretesting procedures.Te intervention lasted fve weeks and consisted of two 30-minute sessions per week.Te sessions included training on a system wall (Kilter, LLC, Boulder, Colorado, United States) attempting a selection of ten boulder problems that were designed to target specifc movement skills (i.e., accuracy, balance, sequencing, technique, arm use, and movement initiation).Te control group, meanwhile, continued with their usual bouldering and training regimens, with an instruction to abstain from using the system wall during the intervention period.

Participants.
Given the absence of comparable studies, a comprehensive power analysis could not be performed.Nevertheless, diligent eforts were made to enlist a maximal number of eligible participants within the local context.Tis involved reaching out to all eligible climbers in the local climbing club and consistently visiting the climbing gym to establish contact with potential participants.Te fnal study population comprised 13 advanced, female boulder climbers (Table 1) who ft the inclusion criteria of having to be active climbers, defned as conducting at least one weekly climbing session, but not having completed a boulder problem graded higher than IRCRA 20 (International Rock-Climbing Research Association; scale ranges from 1 to 32).Tey were also required to not have used system wall climbing as part of their training program before enrolling in the study.Moreover, they had to be free of injuries or illnesses that could limit their ability to participate in the training or exert maximal efort in the testing.Te recruitment phase spanned from September 26 to October 7, 2022, with participants maintaining their involvement in the study until the conclusion of all training and data collection.Te last posttest assessment took place on November 20, 2022.Te study procedures conformed to the latest revision of the Helsinki declaration, followed Norwegian laws and regulations, and were processed by the Norwegian Center for Research Data (reference: 992074) and the local ethics committee (reference: 21/08477-3).
2.3.Procedure.Two boulder problems were set to assess performance and movement skills.Te climbers were given four attempts on each boulder problem with up to one minute rest between attempts on the same problem and two minutes rest between boulder problems.Adapting the approach of a similar study from our lab [16], the attempts on the two system wall boulders were recorded using a stationary iPad (Apple Inc., Cupertino, California, US) positioned approximately four meters away from the center of the wall, ensuring a perpendicular line of sight to the wall.Of the maximum four attempts, the best attempt was stored (i.e., the attempt in which the highest hold was reached).Tree independent experts in the feld with several years of experience as climbing trainers for national and international levels were then engaged to provide feedback on the recordings using a simplifed version of the climber's movement performance assessment tool developed and validated by Taylor and colleagues [10].Te score sheet included six items (accuracy (ACC), balance/ fuidity (BAL), sequencing/exploration (SEQ), technique (TEC), arm position (ARM), and movement initiation (MOV); see Table 2 for more information about each item), each using a fve-point scale ranging from 1 (poor or nonexistent component) to 5 (fawless demonstration of skilled performance).Five random videos were displayed twice to facilitate the calculation of intrarater reliability.Te experts were blinded to the participant's assigned group and unaware of whether they were viewing a pre-or posttest recording.To further ensure anonymity, any visible faces were removed before the footage was presented.Moreover, the experts were externally recruited and had no prior knowledge about the study design or the participants.

2
Translational Sports Medicine Performance was registered as the number of attempts required to complete the boulder problem or as the number of holds reached on their best attempt if they could not complete the boulder problem.In accordance with competition rules, one point was awarded for each handhold successfully grasped and utilized to move upward on the climb, while half a point was given for touching the next handhold, but not being able to sustain it.In addition, the climbing duration (time from the second foot left the foor until both hands were touching the top hold) was collected from the footage by using time stamps in a commercial video player software (VLC Media Player 3.0, VideoLAN, Paris, France).Climbing duration was only analyzed for the eight participants that completed the boulder problems at both pre-and posttest (three in the control group and fve in the experimental group).

Intervention.
Te participants in the experimental group replaced 30 minutes of their regular climbing and training time to performing the movement skill training twice per week for fve weeks, with at least one full day of rest between the two weekly sessions.Te experimental group conducted their initial session under supervision and was presented with a selection of ten boulder problems, systematically increasing in difculty, which they were to engage with on the system wall.Te boulder problems were designed to progressively increase in difculty by incorporating elements such as smaller footholds, more unbalanced positions, longer moves, and intricate sequences, as these would be relevant for the expert rating items.We deliberately minimized increases in difculty that solely relied on reducing hold size, ensuring that technical skills remained more crucial than sheer strength.Subsequent sessions predominantly operated without direct supervision.
To guarantee the participants' compliance with the training protocol and to oversee the progression of their training, the researchers attended the ffth and tenth sessions.Our continuous presence was consistently upheld throughout the intervention period to address any queries or challenges participants encountered during the training.Te participants were instructed to work on four problems in each session, beginning with the easiest four, and to spend no more than fve minutes working on each problem before progressing to the next.If they completed a problem with relative ease, they could omit the problem from the following session and replace it with a more challenging one, while the remaining three from the preceding session were still included.Hence, each session lasted around 30 minutes including rest periods.For the remainder of their time in the gym, the participants climbed and trained as usual on conventional bouldering walls.Te control group continued their regular routines throughout the intervention period.

Statistical Analyses.
Te analyses were performed using the commercial statistical software SPSS (IBM Corp. Released 2020.IBM SPSS Statistics for Windows, Version 27.0.Armonk, NY: IBM Corp).In addition to a Shapiro-Wilk test revealing primarily nonparametric distributions, the ordinal nature of much of the data along with the low number of participants, we opted for nonparametric statistical tests.Tus, between-group diferences were assessed using Mann-Whitney U-tests, whereas within-group diferences from pre-to posttest were addressed with Wilcoxon signedrank tests.Intrarater reliability and interrater reliability were assessed with the intraclass correlation (ICC) of the absolute agreement, which considers both the degree of correlation and the agreement among raters, as well as between ratings, thereby evaluating if the ratings are correlated and also if they are identical in absolute terms.For the interrater reliability, all items were compared across the three raters.For intrarater reliability, ratings of the six items were compared across two ratings of the same video by each rater, and the average ICC for each reviewer was calculated.Te choice to average all items in the reliability analyses was made due to the potential interrelated nature of diferent items (e.g., arm position and movement initiation) and to provide a more substantial value describing the reliability of each rater.Values within the range of 0.21-0.40indicated fair agreement, 0.41-0.60suggested moderate agreement, 0.61-0.80 was indicative of substantial agreement, and 0.81-0.99signifed near-perfect agreement [17].Statistical signifcance was accepted at p < 0.05.Training compliance, number of attempts, and climbing time are displayed using median and range, while the remaining results are presented as means with standard deviations and efect sizes (ESs) for the changes.Te ESs were calculated as the Z-score divided by the square root of the number of observations (n) and interpreted as trivial (<0.10), small (0.10-0.30), medium (>0.30-0.50),and large (>0.50) [18].

Baseline and Training.
Te two groups were not diferent at pretest in any of the collected variables (p � 0.093 − 0.779).Te study participants self-identifed predominantly as boulderers and reported dedicating an Translational Sports Medicine  and 2 for all individual ratings on each boulder problem separately.For the individual data, improvements in expert ratings were generally more pronounced for participants receiving low scores (<2) at pretest, whereas high pretest scores were generally associated with negligible improvements, and in some cases with declines.

Intrarater Reliability.
For the intrarater reliability across all items, the ICC demonstrated near perfect agreement for two of the raters (ICC � 0.894 ± 0.062 and ICC � 0.895 ± 0.060), whereas one exhibited substantial agreement (ICC � 0.807 ± 0.181).Note that all analyses were conducted both with and without the involvement of the third rater (the rater with marginally lower reliability score), and the resulting diferences in the outcomes were negligible and did not afect any of the results.

Interrater Reliability.
Te interrater reliability across the six variables ranged from 0.747 to 0.850 with an average of 0.795 ± 0.039.Te raters reached substantial agreement for ACC (ICC � 0.799), SEQ (ICC � 0.757), ARM (ICC � 0.794), and MOV (ICC � 0.747) and near perfect agreement for BAL (ICC � 0.850) and TEC (ICC � 0.820).Sensitivity analyses with exclusion of the rater with weaker reliability resulted in a slight reduction in the mean ICC (0.786 ± 0.041), further supporting the decision to include all three raters in the analyses.

Bouldering Performance.
For bouldering performance, fve of the seven participants in the experimental group managed to climb boulder problem A in the pretest, whereas all of them completed it in the posttest.Five of the six participants in the control group completed both boulder problems during the pretest and all were able to do it at posttesting.For boulder problem B, all the participants in the experimental group managed the boulder problem within four attempts at both pre-and posttest, whereas the number of tries increased from four to six (all participants) from preto posttest in the control group.
3.6.Attempts.Tere were no changes in the number of attempts to complete the two boulder problems from pre-to posttest in either group (ES � 0.37-0.74,p � 0.317-1.000),and no signifcant diferences were observed between groups at any testing point (ES � 0.05-0.22,p � 0.255-0.904).Initially, the experimental group's median number of attempts for boulder problem A was 1 (range: 1-3), while the control group required a median of 2 (range: 1-3).By the posttest, the experimental group required a median of 1 (range: 1-3), with the control group slightly improving to a median of 1.5 (range: 1-3).For boulder problem B, both groups started with a median of 1 attempt; the experimental group range was 1-2, and the control group range was 1-4.At the posttest, the experimental group remained at a median of 1 (range: 1-3), while the control group needed a median of 1.5 attempts (range: 1-3).

Discussion
Tis study sought to compare the efects of climbing technique practice using a system wall and self-regulated, conventional bouldering training on motor skills among advanced, female boulderers.While our fndings ofer preliminary insights, they also highlight the necessity for studies with larger participant pools to enhance the robustness and generalizability of the results.Both groups exhibited improvements posttraining, with signifcant enhancements in balance and fuidity (BAL) noted only in the experimental group.Importantly, the low number of Translational Sports Medicine participants and the marked individual diferences must be considered when interpreting the results.
Te BAL metric gauges a climber's adeptness at maintaining a seamless fow of movement while avoiding destabilization [10].Te system wall training focused participants on specifc movement patterns, enhancing cognitive engagement to refne motor skills through movement analysis and targeted improvement.Conversely, the control group's BAL ratings did not improve, possibly due to the variable nature of self-regulated training, which may lack the structured practice needed to sharpen specifc skills [11,12].Importantly, this argument warrants further investigation into how recreational climbers structure their self-regulated training.It should also be noted that these fndings are infuenced by two of the participants in the control group receiving lower ratings for BAL at posttest, whereas only one participant in the experimental group decreased in this metric.Interestingly, all three participants who decreased in the BAL rating initially received high scores at pretest (4 and 5), suggesting a potential ceiling efect where higher initial scores limit observable improvements.Future research should remain cautious of ceiling efects in performance measurements, especially when employing scales with upper limits.Such limitations  Mean change is displayed with standard deviations (SDs) and the median change is reported with interquartile ranges (IQRs). 6 Translational Sports Medicine can skew results, particularly when initial scores are high, as they limit the scope for detecting improvements.Tis phenomenon was evident in our pilot study and serves as a critical consideration for the design and interpretation of future experiments.Finally, training context diferences-system wall versus conventional walls used by the experimental and control groups, respectively-could suggest that environmental factors like wall type and hold shapes might afect skill transferability, despite not practicing the exact boulder problems tested [19,20].One of the primary limitations of the current study is its low statistical power, a consequence of the small sample size.Future investigations should aim to include more participants to better detect small, but meaningful efects that could be obscured in smaller samples.However, analyzing individual performance reveals nuances not captured in broader analyses.Participants in the experimental group demonstrated improvements in ACC, BAL, and SEQ for both boulder problems.Tese improvements align with the group-level fndings of enhanced balance and fuidity, as well as sequencing and exploration.In contrast, the control group displayed varied responses, with some participants improving and others showing no change or slight declines.Notably, one control group participant's signifcant improvement infuenced the average change in boulder problem A, likely afecting the overall between-group comparison.Tese fndings indicate that system wall training can enhance technical skills and bouldering performance for advanced female climbers, though variable results and lack of signifcant group diferences underscore the importance of tailoring training to individual preferences and abilities.
Two control group participants scored high (>4) for all rated items on boulder problem B initially but showed notable declines at posttest, signifcantly infuencing both the group's overall change and within-group variation.Tese outcomes could be interpreted to suggest that conventional, self-regulated training may be less efective for climbers already profcient in motor skills, while structured approaches like system wall training could ofer greater benefts by focusing more on specifc motor skills or defciencies in their technique.Variability in training responses may also be infuenced by individual diferences such as climbing experience, learning preferences, and engagement with training.Te substantial variability in individual responses to the interventions is an important takeaway from our fndings.Tis variability suggests that future research should tailor analyses to account for individual diferences, which may reveal more nuanced understandings of the factors infuencing performance gains, especially among advanced, female boulderers.Additionally, flming only two boulder problems, despite providing more data than much existing literature, cannot fully represent all climbing-specifc motor skills.Incorporating a broader range of movements could enrich our understanding of the system wall's impact as a training tool.Given these individual trends and the study's limited statistical power, cautious interpretation is needed, but these insights lay groundwork for further research with larger samples.
Trough the prescription of ten boulder problems featuring a gradual escalation in difculty, the system wall training method may have facilitated errorless learning.Te errorless learning approach to bodily learning involves initiating the practice of a new skill at a lower level of complexity [13], thereby enabling participants to engage with minimal efort while implicitly honing technique and form, which can subsequently be applied to overcoming more demanding challenges.Te fndings further correspond with the concept of resonance, wherein the emergent motor actions resonate with the environmental demands, optimizing the energy expenditure through harmonious coordination [21].In contrast, the self-regulated conventional bouldering likely involved a higher frequency of participants attempting boulder problems close to their skill limits, possibly resulting in a greater occurrence of errors in their movements and solutions.Tis heightened error frequency might have extended the time required for this group to identify efective attractors, adapt individualized solutions, and refne movements perceived as efective [22].Again, further research into the habits of recreational climbers is warranted.
All participants completed both boulder problems within four attempts after the intervention, showing slight improvements for the control group in both problems, and for the experimental group in only boulder problem A. Tis change was accompanied by signifcant SEQ enhancements from pre-to posttest in both groups.Of note, the ES for experimental group (ES � 1.54) was approximately twice as large as for the control group (ES � 0.77), indicating that a higher statistical power likely could have detected betweengroups diferences.Te SEQ metric, assessing the ability to determine optimal movement sequences, likely benefted from participants' familiarity with the problems, revisited fve weeks later.While this recollection might have eased task completion, minor diferences in training intensity and duration between the groups also potentially infuenced the results.
Te fndings regarding minor diferences between groups are comparable to a similar study on motor skill learning conducted in our lab which suggested that, albeit not superior to conventional training approaches, technologically augmented feedback may present a viable alternative which reduces the manual labor required for instructions [16].Importantly, the completion of both boulder problems by all participants highlights a limitation in using such tasks for continuous performance assessment, since a ceiling efect is reached.When climbers are able to successfully complete the problems, assessing further improvements becomes impossible, as the task no longer diferentiates between varying levels of performance.To address this, future studies should explore the use of higher numbers of boulder problems or climbs including more moves, as this could allow for a more nuanced perspective of the performance.
Tis study provides new insights on system wall training versus conventional bouldering but has some limitations.It focused solely on advanced female boulder climbers, limiting generalizability to other demographics like male climbers or Translational Sports Medicine those at diferent skill levels.Te small sample size and the study's location in a small locale also restricted participant diversity.Despite these challenges, the results ofer valuable insights and a basis for future studies.Additionally, the decision to include a third rater with weak reliability could be viewed as a limitation.Importantly, sensitivity analyses showed minimal impact from this decision on the outcomes.Potential group cross-contamination was mitigated by separate training facilities, although completely controlling this in such a social sport is challenging without geographical separation beyond what was feasible for this study.Te fveweek intervention period, while brief, did yield some signifcant skill improvements, suggesting that longer durations and more participants could provide further insights into the sustainability and variability of these enhancements.Future studies could explore longer interventions and more participants to better understand these efects.
Traditional studies on sports performance enhancement typically favor long-term training focused on physical variables like strength and endurance [23][24][25][26].However, linking these improvements to real sport performance remains a challenge.Tis study took a diferent approach by adopting a shorter training period and focusing on specifc climbs to evaluate efectiveness.Tis shift towards sport-specifc assessments provides insights into climbers' skill profciency, underlining the potential benefts of short-term, targeted training interventions for practitioners seeking quick performance gains.More importantly, our fndings display the feasibility for future studies using similar designs, but with more statistical power and longer intervention durations.

Conclusions
Te experimental group's system wall training led to enhancements in balance and fuidity, benefts not seen in the control group.Notably, no signifcant between-groups diferences were detected.Te experiences and fndings from our research should be taken into account when designing future studies on the subject.

Figure 1 :
Figure 1: Overview of the individual expert ratings received on the 1-5 scale.(a) and (b) illustrate changes for the experimental group on boulder problems A and B, respectively, while the changes for the control group on problems A and B are shown in (c) and (d), respectively.Te average change is indicated by white markers and asterisks indicate signifcant changes from pre-to posttest.

Table 1 :
Study sample characteristics by groups and combined presented as mean ± standard deviation.

Table 3 :
Mean and median changes in the expert ratings across the two boulder problems.