Soccer is a typical intermittent team sport characterized by repeated high-intensity activities, including sprints, accelerations, decelerations, and rapid changes of direction (Carling et al., 2012). Due to the frequent occurrence of high-intensity movements during matches, players are required to repeatedly perform such efforts during critical phases, while maintaining a high level of aerobic endurance to support intensity regulation and recovery throughout the game (Harper et al., 2019). Notably, both the duration and repeatability of high-intensity running are strongly correlated with an athlete’s aerobic capacity (VO₂max), with this relationship becoming particularly pronounced during the latter stages of the game—especially in the final 15 min (Gharbi et al., 2015; Grgic et al., 2019). Therefore, the accurate assessment of a player’s aerobic capacity is a cornerstone of effective physical preparation in soccer.

Incremental laboratory-based tests using treadmills or cycling ergometers are widely employed to assess cardiorespiratory fitness. However, these tests are time-consuming, require expensive equipment, and often interfere with athletes’ regular training schedules due to the need for repeated laboratory visits (Bassett and Howley, 2000). As a result, indirect assessment methods have gained increasing attention as practical alternatives (Flouris et al., 2010). To address the limitations of conventional lab-based assessments in athletic contexts, Buchheit and colleagues developed the 30-15 Intermittent Fitness Test (30-15_IFT), a field-based test with greater ecological validity (Buchheit, 2009). The 30-15_IFT is an intermittent, incremental shuttle run test incorporating change-of-direction movements, enabling simultaneous evaluation of aerobic and anaerobic fitness, inter-effort recovery capacity, anaerobic speed reserve, and change-of-direction ability (Jeličić et al., 2020). Recent studies have confirmed the high reliability of the 30-15_IFT across various team sports, including handball, basketball, football, ice hockey, and rugby (Impellizzeri and Marcora, 2009; Jeličić et al., 2020; Grgic et al., 2021). One of the key advantages of the 30-15_IFT lies in its output variable—the maximal intermittent running speed (V_IFT)—which can be used to prescribe individualized high-intensity interval training (HIIT) programs. This feature effectively overcomes the limitations of traditional continuous tests, which often fail to capture the sport-specific demands of match play (Čović et al., 2016). Moreover, HIIT prescriptions based on V_IFT have been shown to significantly reduce inter-individual variability in training intensity within teams (with a coefficient of variation [CV] of approximately 3%), thereby enhancing the homogeneity and standardization of group training sessions (Flouris et al., 2010). This provides a more effective strategy for implementing precise physical conditioning interventions in sports science practice.

Although the 30-15_IFT offers a strong alignment with the physiological and movement demands of team sports, its capacity to predict VO₂max remains significantly different from that of gold standard measurements (ES = 0.84–1.10) (Čović et al., 2016; Jeličić et al., 2020). The estimated VO₂max from the 30-15_IFT (VO₂max-IFT) was calculated using the formula established by Buchheit (2008), which was based on a sample of 59 youth athletes (age, 16.2 ± 2.3 years). The validity of applying such a specific equation universally is questionable, as cardiorespiratory fitness is influenced by a multitude of factors, including sex, ethnicity, training status, and lifestyle (Pandey et al., 2016). This issue of generalizability is compounded by a broader sample bias in the existing literature, which has predominantly focused on professional or youth athletes (Stanković et al., 2021). Compared with youth or professional athletes, collegiate athletes exhibit distinct methodological and practical characteristics. On the one hand, they are typically in a semi-professional state, required to participate in high-level competitions while lacking systematic training, monitoring, and rehabilitation resources. On the other hand, their training load and recovery conditions are often markedly constrained by academic commitments, resulting in greater variability in athletic performance and fitness assessment outcomes (Bozzini et al., 2020). Furthermore, collegiate athletes are generally more biologically mature and no longer display the typical growth and developmental characteristics of youth academy players, a distinction that may influence their fitness adaptations and training responses (Gundersen et al., 2025). In contrast, youth athletes are largely managed within centralized academy systems, while professional athletes benefit from well-established training and medical support structures (McFadden et al., 2023). Therefore, focusing on collegiate athletes not only addresses a critical gap in the existing literature but also provides evidence-based insights for training monitoring and fitness evaluation in both collegiate and semi-professional populations.

Therefore, the primary purpose of this pilot study was to evaluate the reliability and validity of the 30-15_IFT for VO₂max prediction in male collegiate soccer players. A secondary aim was to develop a new, population-specific prediction equation for VO₂max and to directly compare its predictive validity against the Buchheit (2008) formula. We hypothesized that while the 30-15_IFT would be reliable, the new population-specific equation would demonstrate superior accuracy and agreement compared to the general formula.

This pilot study aims to verify the reliability and validity of the 30-15 Intermittent Fitness Test (30-15_IFT) in college soccer players and to explore whether the prediction of aerobic capacity requires different prediction formulas for various groups. The preliminary findings revealed that the 30-15_IFT demonstrates high test-retest reliability (ICC = 0.81–0.92, CV% = 1.43–1.69%) for V_IFT, VO₂max, and HR_max. Although the 30-15_IFT showed a large to very large correlation with the gold-standard continuous treadmill test (CT), its validity is limited because of the bias (ES = 0.96–2.44). Our results also showed that the usefulness of the 30-15_IFT was marginal for all outcome measures. Furthermore, the novel, population-specific equation we developed for collegiate athletes yields a marked reduction in bias and enhanced predictive accuracy. These pilot findings critically highlight the inherent constraints of universal prediction models when applied to specific populations.

Reliability is a critical indicator for evaluating measurement error and is typically categorized into absolute reliability (i.e., the degree of variability in repeated measures for the same individual) and relative reliability (i.e., the consistency of an individual’s rank ordering within a group across repeated assessments) (de Vet et al., 2006; Impellizzeri and Marcora, 2009). Relative reliability is commonly assessed using the intraclass correlation coefficient (ICC), whereas absolute reliability is reflected by the coefficient of variation (CV) and typical error (TE). These metrics are of high practical value in both cross-sectional and longitudinal studies (de Vet et al., 2006). The present results showed high absolute reliability and relative reliability for V_IFT, and higher than the standard (ICC>0.69 and CV<5%) set by Buchheit et al. (2011). No significant differences were observed between the two test trials—indicating stable test outcomes with no evidence of a learning effect (Paravlic et al., 2022). Our study showed similar reliability compared to previous studies targeting various athlete populations, including female basketball players (VO₂max-IFT: ICC = 0.96, CV = 4.83%; V_IFT: ICC = 0.85, CV = 5.99%) (Jeličić et al., 2020), futsal players (V_IFT: ICC = 0.92–0.96, CV = 1.4–1.5%; HR_max-IFT: ICC = 0.90–0.91, CV = 1.3–1.5%) (Valladares-Rodríguez et al., 2017), youth rugby players (V_IFT: ICC = 0.89; HRmax: ICC = 0.96) (Scott et al., 2015), and female soccer players (VO₂max-IFT: ICC = 0.94, CV = 1.6%; HR_max-IFT: ICC = 0.96) (Jeličić et al., 2020). The high reliability of the 30-15_IFT observed in our study makes it comparable to other established field tests for soccer players, such as the widely used Yo-Yo Test, which has demonstrated similar reliability in previous studies (ICC: 0.78–0.98, CV: 3.7%–19.0%) (Grgic et al., 2019). In addition, Buchheit reported a strong correlation between performance on the 30-15_IFT and the Yo-Yo Intermittent Recovery Level 1 (IR1) test (r = 0.75; 90% CI = 0.57–0.86), indicating a degree of convergence between the two assessments. Although these tests may target slightly different physiological capacities, both exhibit similar sensitivity to training-induced changes (Stanković et al., 2021). Taken together, the 30-15_IFT can be regarded as a reliable and reproducible field-based tool for assessing aerobic fitness in collegiate male soccer players.

Although the continuous incremental treadmill test (CT) is widely regarded as the “gold standard” for evaluating aerobic endurance (Poole and Jones, 2017), its practical application is often constrained by environmental conditions, cost, time requirements, and technical complexity. In contrast, the present study established good criterion validity for the 30-15_IFT. The primary outcome measures demonstrated large to very large correlations (r = 0.62–0.77) with their respective counterparts from the continuous treadmill test, supporting the practical utility of the 30-15_IFT for assessing aerobic fitness in collegiate male soccer players. Previous researches also support this opinion. Previous studies have shown similar results. Čović et al. reported moderate-to-strong linear correlations between the 30-15_IFT and CT in both VO₂max (r = 0.67) and HR_max (r = 0.77) in elite female soccer players (Čović et al., 2016). Similar correlations (r = 0.69–0.74) were reported by Jeličić et al. between outcome measures taken during the 30-15_IFT and CT in female basketball players (Jeličić et al., 2020). Notably, while the strong correlations suggest the 30-15_IFT is a valid measurement, the question of whether it truly reflects an athlete’s capacity requires careful consideration. Our findings revealed a notable result: despite the strong relationship, the mean differences between the tests were practically significant, evidenced by moderate to very large effect sizes (ES = 0.96–2.44). Moreover, the Bland-Altman plots showed wide limits of agreement and a visible bias line. It indicated that the outcome measures of the 30-15_IFT have large random error, limiting the utility of the 30-15_IFT as a direct proxy for true aerobic capacity. Previous studies have also reported this significant difference in various populations (female basketball players: ES (d) = 0.84–3.23; female soccer players: ES (d) = 0.98–1.60; Infantry members: ES (η²) = 0.158–0.623) (Čović et al., 2016; Jeličić et al., 2020; Paravlic et al., 2022). On the one hand, previous studies have reported a typical difference of 2–5 km/h between the two tests and have been confirmed in multiple empirical studies (Buchheit et al., 2010; Jeličić et al., 2020). The higher V_IFT values may be attributed to the protocol of the 30-15_IFT, as its intermittent structure and constant 180° changes of direction place significant demands on both anaerobic capacity and the ability to efficiently change direction (Darrall-Jones et al., 2015; Scott et al., 2015). In this study, VO₂max was estimated using Buchheit’s modified formula incorporating V_IFT, age, and body mass. Results showed that the higher V_IFT values contributed to greater VO₂max estimates. Furthermore, Jeličić et al. reported that a familiar training environment may better enable athletes to reach higher speeds compared to laboratory-based continuous treadmill running, leading to an overestimation of VO₂max (Jeličić et al., 2020). In contrast, the lower HR_max-IFT observed during the 30-15_IFT may be linked to the testing environment. The unfamiliar and restrictive nature of laboratory treadmill testing can induce psychological stress and elevate heart rate, whereas the familiar field setting of the 30-15_IFT may elicit a more ecologically valid physiological state (Buchheit, 2008; Paulsen et al., 2023). On the other hand, the observed differences between the two tests may lie in the limitations of applying a single, universal prediction equation to a demographically distinct population. The original Buchheit (2008) formula was developed on a different cohort, which may not accurately capture the specific characteristics of our sample. To investigate this, we tried to develop a population-specific model and directly compared its predictive performance against the Buchheit (2008) equation. The direct comparison of the models revealed that our equation exhibited a substantially lower systematic bias (Bias ≈0 vs. 4.08 mL/kg/min) and a smaller standard error of the estimate (SEE = 2.90 vs. 4.91 mL/kg/min). While this improved accuracy may be partly due to the model being calibrated for our specific and homogenous cohort of male collegiate soccer players, the observed reduction in prediction error highlights a key takeaway. It suggests that applying a general prediction equation to a specific population can lead to significant inaccuracies. Therefore, our findings strongly advocate for the development and use of population-specific models to enhance the precision of field-based physiological assessments.

In addition, the study evaluated the usefulness of the 30-15_IFT by comparing TE versus SWC. This comparison can help coaches conclude the significance of changes in performance due to training interventions or other factors. The results indicated that the usefulness of VO₂max-IFT, HR_max-IFT, and V_IFT was marginal with TE > SWC. However, for V_IFT, TE at 0.35 km/h compared to SWC at 0.27 km/h resulted in a performance change of less than one phase of the exercise (±0.5 km/h). This suggests that an individual’s performance change of less than one phase (±0.5 km/h) can be considered ‘real and meaningful’. These findings suggest that while the 30-15_IFT should not be used as a direct substitute for laboratory-based assessment of maximal aerobic capacity, its high reliability and the usefulness in V_IFT make it an excellent and practical tool for monitoring meaningful longitudinal changes in sport-specific performance and assisting coaches in making informed decisions.

While our study provides valuable insights into the reliability and validity of the 30-15_IFT, several limitations should be acknowledged. First, the reliability was derived from only two trials. Although this is acceptable for calculating the ICC, more robust estimates of practically important values like typical error (TE) and the coefficient of variation (CV) are typically achieved with three or more trials (Hopkins, 2000). Second, our validity analysis, particularly the Bland-Altman plots, revealed wide LoA, suggesting that large random measurement error may limit the interchangeability of the tests for individual assessment. Future studies should consider using larger samples to better quantify this variability. The most significant limitation of this study is the small sample size (n = 20), which restricts the statistical power and generalizability of our findings.

As a pilot study, this work highlights several critical directions for future research. First, the proposed prediction equation requires external validation in a larger, more diverse cohort of collegiate athletes to establish its robustness and generalizability. Second, subsequent research should investigate whether similar predictive biases exist in other distinct populations, such as female collegiate athletes or those from different sports. Finally, future models could incorporate additional variables (e.g., anaerobic speed reserve or change-of-direction metrics) to determine if they further enhance the precision of VO₂max prediction.

Final conclusions must be formulated carefully due to the small sample size. The 30-15 Intermittent Fitness Test is a reliable tool for assessing sport-specific fitness in male collegiate soccer players. Changes in V_IFT of less than one phase of the exercise (±0.5 km/h) are likely to represent a meaningful change in performance. However, due to significant systematic bias and large random error, it cannot be used interchangeably with continuous treadmill testing for assessing maximal aerobic capacity. Furthermore, this study demonstrates that general prediction equations for VO₂max can introduce considerable error when applied to specific populations. Future research should therefore focus on developing and validating these prediction models in larger, more diverse populations to improve their predictive accuracy and generalizability.