Nineteen experienced male university level soccer players who regularly participated in competitive national level soccer tournaments (age: 20.2 ± 1.1 years old, body mass: 66.8 ± 6.7 kg, height: 174 ± 5 cm, and soccer experience: 13.6 ± 2 years) were recruited for the multiple V-cut test. All participants were free from any lower limb injuries and active in university level competitions when they participated in this study during the off-season period. All participants provided written informed consent prior to participation in accordance with the research ethical approval obtained from the institutional research ethics committee.

Three different models of shoes were selected for this study. The properties and features of each shoe are described in Table 1.

The selection criteria for the shoes are as follows: (1) categorized as an indoor soccer/futsal shoes, (2) did not possess midsole construction, (3) possessed similar outsole material and hardness, (4) similar heel-to-toe drop value, (5) available friction coefficient (AFC) differences within ±20%, and (6) mass differences within ±20% and differences below 300 g (Nigg and Enders, 2013; Worobets and Wannop, 2015). These criteria were prerequisite to ensure that all the selected shoes possessed no obvious difference in terms of the shoe construction, and any differences would not provide any obvious advantages/disadvantages during the test conducted in this study.

A cutting course similar to that of the study by Worobets and Wannop (2015) was set in a hardwood indoor flooring facility. Before testing, each participant performed an adequate warm-up that was similar to his usual preparation for a match. Each participant was then asked to maximally complete this cutting course consisting of two 90° cut and 135° V-cut maneuvers within a 4.5 m × 4.5 m area (Figure 1) using the three types of shoes with different sizes. During the test, the order of shoes was randomized (no two consecutive trials using the same shoe). The test was repeated with enough intervals (to avoid the effect of fatigue) between trials to obtain three successful trials (no slipping and colliding with cone during trial) for all the three shoes. A photocell timing gate system (Witty System, Microgate, Italy) was utilized to record the resultant time of all the trials. The distance between the participant front foot and the photocell timing gate sensor during the beginning of the trial was fixed at 0.5 m. All the tests were performed on hardwood indoor flooring facilities.

The FBS of each shoe was measured using a similar method described in the study by Worobets and Wannop (2015). Each shoe was applied with forefoot bending forces ranging between 2 and 18 Nm in 6–9 stages until all shoes reach a maximum bending angle of more than 40°. All data obtained from these trials were plotted on a bending force–angle graph as the bending force was the independent variable, and a regression line was fit at least to a minimum of six data points. The bending stiffness of each shoe was measured three times, and the mean values were computed as the representing values.

In this study, the results obtained for all the three shoes were compared to observe any differences in terms of the bending stiffness properties and resultant running performances during the change of direction run test.

The statistical analyses were performed using an open-source statistical software, PSPP (GNU project, version 1.0.1). A comparative analysis across the three shoes was conducted with one-way ANOVA repeated measures, and Bonferroni post-hoc analysis was applied when required. The statistical significance level was set at p < 0.05 for all the analyses.

The FBS for each shoe and the comparative analysis across the three shoes (i.e., one-way ANOVA repeated measures with Bonferroni post-hoc analysis) are shown in Table 2. It was identified that there was a significant difference between the S2 (0.26 Nm/deg.) and the other two shoes [S1: 0.32 Nm/deg. (p = 0.01) and S3: 0.36 Nm/deg. (p < 0.001)].

The results obtained from the multiple V-cut run performance test are shown in Figure 2. As shown in the figure, there are significant differences among the three shoes for the multiple V-cut performance test [F_(2,170) = 4.60 (p = 0.01)] where the participants performed significantly faster when they used the S2 (4.81 ± 0.3 s: post-hoc p = 0.02) and S3 shoes (4.83 ± 0.3 s: post-hoc p = 0.03) than S1 (4.96 ± 0.3 s) while no significant difference was found between S2 and S3 shoes.

For the FBS, the S1 shoe with a softer outsole material (57 HA) has shown to possess significantly higher bending stiffness (S1: 0.32 Nm/deg., S2: 0.26 Nm/deg.) when compared with the S2 shoe (60 HA). In addition, we also found a significant difference between S2 and S3 shoes (Table 2), where both of them have identical outsole material hardness (60 HA). This finding demonstrated that the hardness of outsole material is not a primary factor to determine the FBS of a shoe. It can be assumed that the stiffness could potentially be modulated by the straight-line groove aspect of the outsoles (Lam et al., 2019). The S1 and S3 shoes do not have aggressive medio-lateral straight-line grooves as S2 shoe. This simplex outsole feature could potentially influence the bending stiffness property of S2. As intended, this feature likely affects its relatively lower bending stiffness when compared with other shoes. The results from this study have demonstrated that the outsole geometrical aspects could have a substantial influence on the FBS of a shoe. In addition, the upper sole material and construction of each shoe could also dictate the FBS properties. Another possible explanation on the lack of influence of the FBS of a shoe on the functional test results is that the FBS of all the shoes selected in this study may be much smaller than the FBS of the participants. If the FBS of a human is larger than the bending stiffness of a shoe, then the total FBS will be dominated by the human foot, thus limiting the influence of the FBS of shoes during the functional test. A similar finding was reported in the literature when it was found that if the FBS of a human is larger than the bending stiffness of the running shoes, then any variations in the bending stiffness of the shoes are unlikely to have a significant influence on running performance (Oleson et al., 2005).

It was generally accepted that higher FBS is known to benefit athletes in running performance (Stefanyshyn and Fusco, 2004; Tinoco et al., 2010; Worobets and Wannop, 2015; Day and Hahn, 2020). The studies by Worobets and Wannop (2015) and Day and Hahn (2020) have reported the FBS values (0.10–0.29 Nm/deg. and 0.22–0.33 Nm/deg., respectively) by using a similar method performed in this study. Both studies reported that higher FBS significantly improved running performances. In this study, however, the differences in FBS of the tested shoes do not seem to play a dominant role for running performance because the shoe with the lowest FBS (S2) did not suffer any detriments on resultant performance in change of direction results. The contradicting results found in this study may be due to the fact that these two previous studies (Worobets and Wannop, 2015; Day and Hahn, 2020) applied systematic ways to solely alter the shoe bending stiffness. Under this circumstance, only the FBS property of the tested shoes was altered while retaining the shoe property identical. On the contrary, in this study, an attempt was made to compare commercially available shoes across different models but with similar features and construction. Based on this evidence, it can be speculated that that there could be other performance limiting factors rather than FBS which have influenced the test outcomes during the multiple V-cut change of direction.

The previous study has verified that increasing the shoe mass would predictably reduce running performance from the energetic point of view (Hoogkamer et al., 2016). From the perspective of sprint and cutting performances, the effect of shoe mass seems to have some threshold. The study by Nigg and Enders (2013) on sports footwear found that only shoe mass becomes an important factor for the weight difference of above 300 g in running performance. This finding has been supported by other similar studies. Recent study by Köse (2018) verified that the weight difference of 285 g between two shoes has shown a significant effect on 10-m sprint performance. In another study on basketball shoe, it was reported that the shoe within 20% of weight difference did not significantly influence the sprint and cutting performances (Worobets and Wannop, 2015). Since the three shoes used in this study are within 20% of weight differences (below 285 g of weight differences), it was concluded that the influence of shoe mass is less than small or non-existent due to small weight differences.

The three shoes selected for this study possess relatively small heel-to-toe drop differences (7–10 mm) when compared with one another. While it could not be identified precisely how this small difference could influence the study outcomes, the only possibility that exist is the shoe with higher heel-to-toe drop could possibly offer a slight extra cushioning feature to the shoe. However, the study by Lam et al. (2017) identified that cushioning provides no advantage to the horizontal ground reaction force component that is crucial during short-exertion, high-intensity movement such as short-distance sprint and change of direction tasks. Therefore, it can be suggested that the influence of heel-to-toe drop difference was minimal in this study.

The three shoes selected in this study were also selected in other previous study (Ismail et al., 2020), where the available traction (the AFC) of all the shoes were mechanically measured. The influence of AFC on change of direction and perceived traction performances was observed in the study by Ismail et al. (2020). It was reported that AFC possessed substantial influence on the change of direction and perceived traction performances. Similarly, as reported in this study, Ismail et al. (2020) reported that participants have performed significantly better when using S2 (AFC: 1.34) and S3 (AFC: 1.30) shoes as compared with S1 shoe (AFC: 1.25). Thus, it was suggested that differences on the AFC component between the shoes of S1 and S2 as well as S1 and S3 could potentially influence the outcomes of the study. Currently, there is no existing study that has compared the influence of both AFC and FBS on change of direction performance. Therefore, it is still difficult to establish a clear conclusion, but as observed in this study and previous study (Ismail et al., 2020) it can be speculated that AFC could potentially possess a much dominant influence on change of direction performance as compared with FBS.

This study focused only on the commercially available futsal shoes to provide more practical, user-friendly information. Although careful selection criteria to pick three different futsal shoes representing similar features were made, several factors, including outsole groove patterns and shoe upper materials and construction, were not systematically controlled. Therefore, the results observed in this study might not be generalized for all types of futsal shoes. More testing on various futsal shoe model is warranted. In addition, we tested the shoe with only one playing surface (hardwood). The tests on different types of futsal playing surfaces (e.g., vinyl, plastic-based, or rubberized material) should be warranted in future studies.