The foot strike pattern performed during running is an important variable for runners, performance practitioners, and industry specialists. Versatile, wearable sensors may provide foot strike information while encouraging the collection of diverse information during ecological running. The purpose of the current study was to predict foot strike angle and classify foot strike pattern from Loadsol^TM wearable pressure insoles using three machine learning techniques (multiple linear regression―MR, conditional inference tree―TREE, and random forest―FRST). Model performance was assessed using three-dimensional kinematics as a ground-truth measure. The prediction-model accuracy was similar for the regression, inference tree, and random forest models (RMSE: MR = 5.16°, TREE = 4.85°, FRST = 3.65°; MAPE: MR = 0.32°, TREE = 0.45°, FRST = 0.33°), though the regression and random forest models boasted lower maximum precision (13.75° and 14.3°, respectively) than the inference tree (19.02°). The classification performance was above 90% for all models (MR = 90.4%, TREE = 93.9%, and FRST = 94.1%). There was an increased tendency to misclassify mid foot strike patterns in all models, which may be improved with the inclusion of more mid foot steps during model training. Ultimately, wearable pressure insoles in combination with simple machine learning techniques can be used to predict and classify a runner’s foot strike with sufficient accuracy.

Trail running is an increasingly popular discipline, especially over long-distance races (>42.195 km). Pacing strategy, i.e., how athletes modulate running speed for managing their energies during a race, appears to have a significant impact on overall performance. The aims of this study were to investigate whether performance level, terrain (i.e., uphill or downhill) and race stage affect pacing strategy and whether any interactions between these factors are evident. Race data from four race courses, with multiple editions (total races = 16), were retrieved from their respective events websites. A linear mixed effect model was applied to the full dataset, as well as to two subgroups of the top 10 male and female finishers, to assess potential differences in pacing strategy (i.e., investigated in terms of relative speed). Better finishers (i.e., athletes ranking in the best positions) tend to run downhill sections at higher relative speeds and uphill sections at lower relative speeds than slower counterparts (p < 0.001). In the later race stages, the relative speed decrease is larger in downhill sections than in uphill ones (p < 0.001) and in downhill sections, slower finishers perform systematically worse than faster ones, but the performance difference (i.e., between slower and faster finishers) becomes significantly larger in the later race stages (p < 0.001). Among elite athletes, no difference in pacing strategy between faster and slower finishers was found (p > 0.05). Both men (p < 0.001) and women (p < 0.001), in the later race stages, slow down more in downhill sections than in uphill ones. Moreover, elite women tend to slow down more than men (p < 0.001) in the later race stages, regardless of the terrain, in contrast to previous studies focusing on road ultramarathons. In conclusion, running downhill sections at higher relative speeds, most likely due to less accentuated fatigue effects, as well as minimizing performance decrease in the later race stages in downhill sections, appears to be a hallmark of the better finishers.

Trail running is an emerging discipline with few studies performed in ecological conditions. The aim of this work was to investigate if and how biomechanics differ between more proficient (MP) and less proficient (LP) trail runners. Twenty participants (10 F) were recruited for a 9.1 km trail running time trial wearing inertial sensors. The MP athletes group was composed of the fastest five men and the fastest five women. Group differences in spatiotemporal parameters and leg stiffness were tested with the Mann–Whitney U-test. Group differences in joint angles were tested with statistic parametric mapping. The finish time was 51.1 ± 6.3 min for the MP athletes and 60.0 ± 5.5 min for the LP athletes (p < 0.05). Uphill sections: The MP athletes expressed a tendency to higher speed that was not significant (p > 0.05), achieved by combining higher step frequency and higher step length. They showed a tendency to shorter contact time, lower duty factor and longer flight time that was not significant (p > 0.05) as well as significantly lower knee flexion during the stance phase (p < 0.05). Downhill sections: The MP athletes achieved significantly higher speed (p < 0.05) through higher step length only. They showed significantly higher knee and hip flexion during the swing phase as well as higher trunk rotation and shoulder flexion during the stance phase (p < 0.05). No differences were found with respect to leg stiffness in the uphill or downhill sections (p > 0.05). In the uphill sections, the results suggest lower energy absorption and more favorable net mechanical work at the knee joint for the MP athletes. In the downhill sections, the results suggest that the more efficient motion of the swing leg in the MP athletes could increase momentum in the forward direction and full body center of mass’ velocity at toe off, thus optimizing the propulsion phase.

Various methods for measuring ground reaction forces during ski-jumping have been developed and applied. The main limitations of the current systems are the restriction to the measurement of single phases (parts of the in-run and take-off) and single force components (vertical component only). Therefore, the purpose of this study was to assess a recently developed force measurement system (JH-SJ), which records forces in the vertical (compression and tension) and horizontal (anterior and posterior) directions over the entire hill jumping sequence. Comparing the values between the JH-SJ and calibrated reference systems in a static and ski-jumping-specific condition provided a means of validating the use of JH-SJ. Hill jumps were performed to evaluate the system regarding the feasibility of application and plausibility of the obtained data. Mean absolute difference during static validity evaluation was 3.5 N (2.7%). Root mean square errors during ski-jumping-specific validity evaluation were less than 11.9 N. Vertical forces acted as compression during all phases of the hill jump except for the flight phase, when tension forces were also recorded. Horizontal force acted as friction on the system during in-run, take-off and landing. The small deviation of the measurement system compared to the reference systems suggests high validity of the JH-SJ measurement system. Qualitative analysis of the forces during hill jumps delivered plausible results for all phases, similar to previous studies. Based on this study, the JH-SJ system measures forces with a high degree of accuracy and precision in real conditions (hill jumps), while contributing new insights into ski-jumping kinetics.