Tactical learning in young footballers is typically studied from two dominant perspectives, the cognitive experimental and the ecological, which often overlook the explicit recognition and self-evaluation of taught content (Abernethy et al., 2005; Gumusdag et al., 2025; Zoudji et al., 2010). (2) The ecological approach, which emphasizes adaptive regulation within the player-environment interaction (Araújo et al., 2006; Chow and Kee, 2022). Yet both paradigms frequently neglect a crucial dimension: the player's ability to explicitly recognize and assess the tactical content introduced during training, and to relate it to their own performance. This reflective capacity constitutes a key indicator of meaningful internalization.

Recent studies, in sport pedagogy, suggest that pedagogical that Tactical Games for Understanding (TGfU) approach encourage metacognitive engagement in young athletes (Light, 2004; Harvey and Jarrett, 2014). These behaviors can take the form of verbalisation, explicit recognition of training-related action, or introspective judgements about one's own decisions during play. They reflect a form of reflective judgement, in which players evaluate their choices in light of the tactical principles discussed during prior instruction. Explicit recognition and brief self-evaluation mark a shift from mere execution to the conscious appropriation of tactical principles (Light, 2004; Harvey and Jarrett, 2014). This process reflects a deeper level of integration, as it requires the activation of structured declarative memory. Findings from developmental psychology support this trajectory. Among the youngest age group (U-13), verbalisation tends to be fragmented, intuitive, and unstable, whereas in adolescence (U-15 and U-18), players progressively develop the ability to identify, label, and reapply tactical concepts in gameplay, demonstrating of increasingly structured metacognition abilities, whether implicit or explicit. Recognition and reflective judgement may support the consolidation and transfer of tactical principles, highlights the limitations of behavioral observation alone as an indicator of understanding. This is the rationale underpinning our approach, which introduces an evaluation method centered on recognition and declarative metacognitive judgement. The temporal stability of these judgements will be examined in the next section.

If recognition and reflective judgement serve as indicators of the conscious appropriation of knowledge, they must demonstrate temporal stability to reflect cognitive anchoring. In other words, a player who identifies a tactical action as previously taught should be able to sustain that judgement under comparable, repeated conditions, regardless of fatigue, random variation, or contextual influences. The intra-individual stability of metacognitive judgements thus constitutes a criterion for assessing the durability of learning. This concept, grounded in research from cognitive psychology, rests on the assumption that consistent short-term self-assessments are positively associated with stronger long-term memory representations (Flavell, 1979; Veenman, 2017; Dunlosky and Metcalfe, 2009). Stable judgements across comparable contexts suggest that the learner is reliably accessing declarative knowledge. However, within the domain of sport, this concept remains largely underutilized. Most assessment protocols still rely on immediate performance outcomes or observational data, without accounting for the reliability or short-term stability of self-reported judgements. Recent studies in educational science and sport didactics have shown that immediate metacognitive judgements so-called ‘hot' judgements are often subject to overestimation. They are often influenced by perceived success, familiarity with the context, or the subjective ease of processing (Flavell, 1979; Frank and Kuhlmann, 2017). These judgements are shaped by both immediate internal experiences and pre-existing beliefs about learning, which can result in a significant discrepancy between subjective perception and actual knowledge. This tendency is particularly pronounced in younger children, whose metacognitive development is still in progress. To address this, we propose the use of an intra-session double-testing method. Accordingly, double-pass (DP) does not measure an additional recognition attempt, but the stability of the judgement across repeated exposure. This involves presenting the same tactical situation twice, at short intervals, without informing the players in advance. This implicit protocol enables the measurement of declarative stability in recognition: a player who consistently identifies the same action as having been taught is likely drawing on a stronger and more organized memory than one whose judgement varies. Accordingly, the DP condition should be interpreted as an index of intra-session stability rather than a second recognition performance. This approach aligns with developmental models (Veenman et al., 2006), which suggest that self-regulation, verbalisation, and evaluative accuracy improve with age. Based on this framework, we expect a marked increase in metacognitive stability between the U-13, U-15, and U-18 age groups. These categories mark key developmental stages within academy pathways, where selection, role differentiation, and preparation for senior competition typically occur. Maturation-related factors may, in turn, influence both the recognition of tactical content and the short-term stability of metacognitive judgements (Deprez et al., 2015; Höner and Feichtinger, 2016; Reilly et al., 2000). By combining reflective recognition with its temporal stability, this study aims to deepen our understanding of how tactical knowledge is internalized during training. The DP method provides a novel way of assessing not only what players recognize as having been taught, but also the extent to which their judgements are reliable, structured, and transferable to game contexts.

In light of the theoretical contributions discussed, it is essential to employ a protocol capable of capturing both explicit recognition of taught tactical content and the temporal stability of that recognition. The literature suggests that reflective judgement and stability are complementary dimensions in assessing the depth of learning particularly in intensive training environments, where cognitive load is high and verbalisation is seldom used. To address these demands, the present study adopts an experimental design combining two assessment procedures: a single-pass test (SP) measuring immediate recognition of tactical actions, and a DP test evaluating the intra-session stability of recognition judgements.

This original protocol thus moves beyond single-point assessments by examining the internal consistency of metacognitive judgements over a short time interval. By applying this approach across three distinct age categories (U-13, U-15, U-18), the study aims to assess the influence of cognitive development on the declarative consolidation of tactical learning. The methodology rests on two core hypotheses: (1) older players will demonstrate a greater ability to explicitly recognize previously taught actions; and (2) such recognition will be more stable over time, reflecting stronger memory anchoring and more advanced metacognitive regulation. The following section details the methodological choices made to test these hypotheses, including sample selection, testing procedures, and the development of assessment materials.

To operationalise these dimensions, we developed a structured taxonomy of explicitly teachable tactical actions aligned with the ten core FUT-SAT principles and supported by a systematic literature review. The methodological procedure underpinning its construction and validation is presented below. A systematic review was conducted in accordance with PRISMA guidelines (Moher et al., 2009), covering publications from December 2000 to January 2024. The literature search was conducted in the Web of Science database using a combination of keywords, topic fields, and Boolean operators. The search strategy included the following terms: TI (“football tactical actions”) OR TS (“tactical principles” OR “tactical decision-making” OR “game tactics” OR “tactical knowledge”). Inclusion criteria were: (a) be published in peer-reviewed journals; (b) explicitly reference at least one FUT-SAT principle; (c) concern players in the U-13 to U-18 age categories; and (d) be available in English, French, or Spanish. Studies were excluded if they focused solely on performance outcomes without a tactical rationale, lacked methodological or conceptual rigor, or were published before 2000 (except seminal contributions). Of 237 records initially screened, 25 met all inclusion criteria (Figure 1). Tactical actions were retained if they met three criteria: empirical or theoretical support; a decision-making component; and a contribution to collective organization. Each action had to be clearly formulated, appear in at least one included study, and be defined as a deliberate spatiotemporal adaptation involving intentional decisions on movement, positioning, or interaction in response to dynamic constraints (González-Víllora et al., 2015). Actions were classified according to the four tactical levels proposed by (Costa et al. 2011): general, fundamental, operational, and specific. Each action also required observable coordination with at least one teammate toward a collective objective (O'Connor et al., 2018). Frequency of occurrence and execution complexity were excluded as selection parameters due to high variability across players and contexts (Mackenzie and Cushion, 2013). Likewise, motor, cognitive, and contextual complexity could not be systematically evaluated within a purely declarative framework. An initial set of 47 tactical actions was identified and mapped to corresponding FUT-SAT principles. Two UEFA A-licensed or UEFA Pro-licensed coaches specializing in youth development independently evaluated each action using descriptor sheets comprising the action name, a concise definition, the associated principle, and evaluation criteria. They assessed definitional clarity, alignment with the corresponding principle, and compliance with inclusion criteria, and could propose rewording, substitution, or exclusion. Feedback was collected anonymously via an online form and re-evaluated in a second validation round using a Delphi-style consensus procedure (Hsu and Sandford, 2007; Quartiroli, 2025). Following this process, 40 actions were retained; four were excluded for insufficient specificity and three for redundancy. Inter-rater agreement reached 85% (Cohen's κ = 0.78), indicating substantial agreement. The final list appears in Table 1.

The assessment material consisted of 40 tactical actions representative of commonly trained situations, based on the taxonomy developed by (Acoumambo and Zoudji 2023). Each action was rendered as a 2D animation using TacticalPad software (45° oblique angle, 12 frames per second) and validated by two UEFA Pro License coaches. Each animated sequence was followed by five response statements, one of which was a declarative recognition item: “I know this action; it was covered during my training” (Figure 2). In the SP condition, each action was presented only once. In the SP condition, each action was presented only once. In the DP condition, each action appeared twice after a pseudo-random delay of 7 to 15 intervening items (cf. Zoudji and Thon, 2003) (Figure 3).

Sessions were held in an auditorium between May and June 2024, with groups of between 16-22 players. Each session comprised three phases: (1) a brief demographic questionnaire (age, years of practice; Table 1); (2) three practice trials to familiarize players with the response format; and (3) the main sequence of tactical actions. The experimental timeline followed a fixed structure: each animation lasted approximately 6- s, followed by an 8-s response window. An 8-s response window, calibrated in pilot tests with U-13-U-18, balanced reading/decision time with temporal continuity and was applied identically in SP and DP. Inter-item intervals: 1-2 s. Repeated items in the DP condition were reintroduced after 7-15 intervening trials (pseudo-random). All repeated stimuli were identical, ensuring so DP scores index judgement stability rather than additional perceptual learning. Each animation was projected onto a 3 × 4 m screen. At the end of each video, the final frame was held on screen, and players had eight seconds to respond using an interactive keypad system (QuizzBox). Responses could be modified within this time window. The order of presentation of all stimuli was fully randomized, and players were not informed of any repetitions. The lack of motor engagement and brief stimuli limited practice and fatigue effects. No order effects were observed, further supporting the internal validity of the protocol. All responses were automatically recorded in an Excel file. Content validity was supported by expert review (Cohen's κ = 0.78); procedural validity was reinforced by the DP index of intra-session judgement stability.

Mean scores (expressed as percentages) revealed a decline in declarative stability score between the SP and DP conditions across all age groups (Table 3). This decrease was most pronounced among the youngest participants: U-13 (Δ = 12.87), U-15 (Δ = 11.01), and U-18 (Δ = 9.59 percentage points). Shapiro-Wilk tests indicated acceptable approximation to normality at the group level (all p>0.050). Visual inspection of Q-Q plots and residual distributions revealed no substantial deviations. Homogeneity of variances was supported (Levene's tests, all p> 0.05). Based on these diagnostics, ANOVA was retained as an appropriate analytical approach.

Inter-group comparisons (age categories): In the SP condition, only the comparison between U-13 and U-18 players was statistically significant (t (148) = −2.92, p = 0.004, d = −0.60, 95% CI [-0.97,−0.19]). In the DP condition, this difference was more pronounced (t (148) = −3.83, p < 0.001, d = −0.78, 95% CI [-1.17,−0.39]). Intra-group comparisons (SP vs DP): Significant differences between SP and DP conditions were observed within each age group: U-13: t(74) = 7.15, p < 0.001, d = 1.04, 95% CI [0.67, 1.41]; U-15: t (74) = 4.15, p < 0.001, d = 0.60, 95% CI [0.26, 0.95]; U-18: t (74) = 3.58, p < 0.001, d = 0.52, 95% CI [0.18, 0.86]. Across all groups, SP scores were higher than DP scores.

The ANOVA revealed a significant main effect of age group, F (2, 222) = 9.27, p = 0.001, η2P = 0.05, 95% CI [0.01, 0.08], and a strong main effect of test condition, F (1, 222) = 673.33, p < 0.001, η2P = 0.75, 95% CI [0.71, 0.78]. A significant interaction between age group and test condition was also found, F (2, 222) = 4.87, p = 0.008, η2P = 0.042, 95% CI [0.003,0.07]. These results indicate statistical differences between age groups, test conditions, and their interaction (Figure 4).

Moderate threshold (≥ 0.50). The data show an age-related increase in the number of actions recognized as having been addressed in the SP condition. In the DP condition, however, stability was low or even absent among the youngest players. McNemar tests confirmed this trend: U-13 (p = 0.001), U-15 (p = 0.001), U-18 (p = 0.20) (Table 4). High threshold (≥ 0.80). At this more stringent threshold, no action was judged to be fully integrated, regardless of age group or test condition (McNemar: p > 0.05 for all groups) (Table 4).

When examined across all 40 actions, stability remained partial overall, with consistently higher scores in the SP condition. U-13 players stabilized an average of 18 actions in SP (45%) and 13 in DP (33%). U-15 players stabilized 20 in SP (51%) and 16 in DP (40%). U-18 players reached 22 in SP (55%) and 18 in DP (45%) (Table 4 and Figure 5). In summary, these results suggest that: (1) immediate recognition (SP) tends to produce an overestimation of learning; (2) judgement stability (DP) reflects only partial consolidation, even among U-18 players; and (3) the overall number of genuinely stabilized tactical actions remains low, raising concerns about the pedagogical effectiveness of current training practices in elite youth development.

The present findings support the relevance of a metacognitive approach to evaluating tactical learning in youth footballers. Across all age groups, scores were consistently lower in the DP condition compared to the SP condition, indicating that immediate judgements are less stable when reassessed within the same session. This difference may partly reflect overconfidence in single-pass responses, but should be interpreted cautiously since the design does not directly measure learning magnitude. Importantly, the discrepancy between SP and DP scores was more pronounced in younger players, suggesting that it may reflect age-related differences in verbalisation, attentional control, and judgement calibration rather than genuine disparities in tactical competence. These interpretations are in line with developmental research on metacognitive variability and short-term judgement reliability. The comparatively lower but more consistent scores observed in the DP condition likely may capture a more conservative estimate of declarative accessibility. This finding is consistent with research on judgement stability in short temporal windows (e.g., Serra and Dunlosky, 2010), indicating that repeated assessments can differentiate intuitive recognition from more robust forms of declarative anchoring. Nevertheless, further research is required to determine whether DP performance predicts long-term retention or in-game decision-making, which were not assessed in the present study.

The age-related differences observed in this study suggest a gradual improvement in the ability to recognize and produce more consistent judgements of taught tactical content with increasing cognitive maturity. U-18 players obtained higher scores than U-13 players in both SP and DP conditions, suggesting more advanced declarative accessibility and judgement consistency. These patterns are compatible with developmental models describing the progressive emergence of explicit metacognitive abilities during adolescence (Flavell, 1979; Veenman, 2017). Among U-13 players, greater variability in the DP condition may reflect less stable verbalisation, weaker attentional control, and less consistent judgement calibration at earlier developmental stages. Importantly, these results do not imply lower tactical potential, but rather point to differences in how young players monitor and articulate their knowledge. Taken together, these findings suggest that instructional demands and assessment formats may need to be adapted to learners' cognitive maturity, particularly in the youngest age groups.

A notable finding is the relatively modest DP performance, even among U-18 players. No age group reached the more stringent threshold of stable recognition (≥ 0.80), and the highest mean values remained below half of the presented actions. Moreover, no signs of attentional disengagement were observed throughout the trials, as indicated by the absence of a significant order effect between passes. This behavioral stability suggests that the double-pass procedure did not induce any notable cognitive fatigue and that participants remained consistently engaged throughout the entire task. While these results do not imply insufficient learning, they suggest that only a limited proportion of taught tactical content is consolidated at a declarative and stable level. Although the intra-session stability observed in this study does not directly demonstrate long-term consolidation, it may be considered a preliminary indicator of cognitive structuring. In other words, increased stability within the same session could reflect an emerging organization of declarative knowledge that may later support durable retention. However, this assumption remains to be empirically tested through longitudinal protocols specifically designed to assess consolidation over time. Several hypotheses may account for this pattern. First, at this developmental stage, a substantial share of tactical knowledge may be integrated implicitly and mobilized in action without explicit verbal recall. Second, current training environments may place greater emphasis on procedural execution rather than reflective verbalisation, potentially constraining the consolidation of explicit declarative representations. Taken together, these observations indicate that pedagogical approaches incorporating more regular metacognitive dialogue, guided reflection, and explicit reference to tactical principles may support deeper declarative anchoring over time. These interpretations should nonetheless be considered with caution, as multiple contextual variables such as coaching methodology, training volume, and seasonal phase were not controlled in the present design and may have influenced the results.

Several pedagogical recommendations emerge from this study: (1) Incorporate regular opportunities for reflective verbalisation such as self-assessment tasks, post-session debriefings, or metacognitive questionnaires to support the stability and explicit evaluation of learning. (2) Align instructional strategies with players' developmental stages: for U-13s, prioritize visual aids and guided tactical games; for U-18s, emphasize tasks involving transfer, self-explanation, and autonomous decision-making. (3) Employ repeated assessment procedures to distinguish between immediate recognition and lasting consolidation, and avoid basing instructional adaptations solely on early or surface-level judgements. (4) Finally, this study appears to support the value of a pedagogy grounded in deep learning and applied metacognition, aimed at fostering tactical knowledge that is explicit, stable, and transferable beyond the immediate context (Hattie and Donoghue, 2016).

This study presents several limitations. First, recognition was assessed solely through a declarative recognition task. Complementary methods such as open-ended verbalisation, qualitative interviews, or performance-based assessments could enrich the understanding of how tactical knowledge is expressed. Although the recognition-based format provides a controlled way to probe declarative anchoring, the absence of external validation and diagnostic accuracy measures (e.g., hits, false alarms, d′) represents a methodological limitation of the present design. From a validity standpoint, however, the task employed in this study closely aligns with the cognitive demands of real-game situations, thereby supporting its content validity. Furthermore, the correspondence between metacognitive judgements and actual performance outcomes suggests satisfactory criterion validity. These findings are consistent with studies grounded in signal detection theory applied to sport performance (e.g., Oudejans and Pijpers, 2009; Tenenbaum et al., 2013) which emphasize the relevance of approaches sensitive to attentional and decision-making fluctuations in dynamic environments. Future studies could include coach-verified exposure logs and optional tactical justifications to enhance construct validity. Second, although the experimental protocol was rigorous, it was conducted outside the natural training environment. Further ecological validation in real-game contexts is required to confirm the applicability of the findings. Contextual and training-related variables (e.g., coaching approach, seasonal period) were not controlled in the current design. These factors may substantially influence recognition stability and should be treated as covariates or control variables in future research. Third, the study also did not control for instructional quality, including coaching style and the frequency of tactical reinforcement. Future research should explore the relationship between the stability of metacognitive judgements and actual in-game performance. It would also be valuable to examine the longitudinal effects of metacognitive feedback mechanisms implemented during training. Despite these limitations, this work makes an original contribution by articulating the metacognitive dimensions of tactical stability within a controlled experimental framework. Theoretically, it advances our understanding of the stability of declarative accessibility from a metacognitive perspective. Methodologically, it proposes a reproducible protocol for identifying explicit stability indicators. Practically, it provides concrete avenues for designing training programmes that integrate self-regulation and metacognitive feedback into the development of young players.