*3.1. Mean Value Changes in Vertical Oscillation, Normalized Frequency, and Duty Factor*

The rmANOVA value for osc\_CoM showed a non-significant main effect of the factor expertise level (*p* = 0.33, *η* 2 *<sup>p</sup>* = 0.189) and a non-significant interaction effect of expertise level and speed, with a high effect size (*p* = 0.576, *η* 2 *<sup>p</sup>* = 0.142). The main effect of the factor speed was also not significant (*p* > 0.999, *η* 2 *<sup>p</sup>* = 0.054).

The rmANOVA for SF\_norm showed a significant main effect of the factor speed, with a high effect size (*p* = 0.018, *η* 2 *<sup>p</sup>* = 0.438) and a non-significant interaction between expertise level and speed (*p* > 0.999, *η* 2 *<sup>p</sup>* = 0.001). The main effect of the factor expertise level was not significant (*p* > 0.999, *η* 2 *<sup>p</sup>* = 0.014). In both groups, SF\_norm increased from 10 to 15 km/h.

The rmANOVA for DF showed significant main effects of expertise level (*p* = 0.018, *η* 2 *<sup>p</sup>* = 0.502) and speed (*p* = 0.018, *η* 2 *<sup>p</sup>* = 0.908), with high effect sizes. However, the interaction effect of expertise level and speed was not significant (*p* > 0.999, *η* 2 *<sup>p</sup>* = 0.008). In summary, DF decreased from 10 to 15 km/h and was overall higher for NOV than for EXP.

#### *3.2. Changes in Variability according to the Expertise Level at Two Running Speeds*

With regard to the CV of osc\_CoM, rmANOVA showed significant main effects of expertise level (*p* = 0.018, *η* 2 *<sup>p</sup>*= 0.792) and speed (*p* = 0.018, *η* 2 *<sup>p</sup>* = 0.408), as well as nonsignificant interaction effects (*p* = 0.084, *η* 2 *<sup>p</sup>* = 0.279), each with high effect sizes. This implies that the CV of osc\_CoM was higher for NOV than for EXP. The CV of osc\_CoM increased with an increase in speed.

The rmANOVA for the CV of SF\_norm showed a significant effect of the factor expertise level, with a high effect size (*p* = 0.018, *η* 2 *<sup>p</sup>* = 0.435). However, the effect of speed (*p* > 0.999, *η* 2 *<sup>p</sup>* < 0.001) and the interaction effect of expertise level and speed (*p* > 0.999, *η* 2 *<sup>p</sup>* = 0.025) were not significant. Accordingly, the results showed that the CV of SF\_norm was higher for NOV than for EXP.

The rmANOVA for the CV of DF had no significant effects, but a high effect size was found for the factor expertise level (*p* = 0.520, *η* 2 *<sup>p</sup>* = 0.155). In contrast, the effect sizes for speed (*p* > 0.999, *η* 2 *<sup>p</sup>* = 0.017) and the interaction effect of expertise level and speed (*p* > 0.999, *η* 2 *<sup>p</sup>* < 0.001) were low.

#### **4. Discussion**

The aim of this study was to analyze the effects of the expertise level on key biomechanical parameters and the variability of these parameters at two different running speeds. It was hypothesized that regardless of the running speed, expert runners are characterized by lower mean values and lower variability of all the considered parameters compared with their novice counterparts. The results indicated that the expert runners had a lower duty factor than the novices. Furthermore, the experts showed a significantly lower variability than the novices with regard to vertical oscillation of CoM and normalized stride frequency, independently of the running speed, but no differences in variability were observed for the duty factor. Based on the findings on this study, our hypotheses can be only partially accepted.

#### *4.1. Lower Duty Factor for Expert Runners*

EXP and NOV did not differ significantly in terms of osc\_CoM (EXP, 10 km/h: 91.43 mm, 15 km/h: 95.06 mm; NOV, 10 km/h: 82.3 mm, 15 km/h: 81.36 mm). These findings are interesting, since more experienced runners are expected to have a better running economy [9,20], which was shown to be associated with a lower osc\_CoM [2,8,9].

With regard to the results for SF\_norm, they are in line with those of other studies [13,43], that is, an increased cadence was observed with an increase in speed. Based on these results, it can be suggested that SF\_norm is not directly affected by the expertise level but, rather, is a function of the running speed.

DF, which describes the step time relative to the flight time, decreased with increasing speed in both groups in this study, and this finding is also in line with other published studies [8,44]. Furthermore, EXP showed an overall lower DF than NOV (10 km/h: 17.3% less; 15 km/h: 20.3% less); this indicates that EXP had a longer flight phase than NOV at a given stance time. The interpretation of the DF results is not straightforward, but an optimal level of DF seems to exist for runners. Even though a lower DF was shown to be associated with better running economy, a DF value that is too low could be uneconomical, given the high muscle activation that occurs over a very short stance time. Further, a DF value that is too high may indicate high start–stop accelerations and, therefore, a waste of mechanical work [45].
