Corrected link here: doi.org/10.1016/j.jb...

⚡🚫 Caution: Be cautions when using variability-based metrics based on markerless data. The higher inter-trial variability could potentially lead to misleading results.

✅ 📉 Mitigation Strategies: Do not trust single waveforms, use avareged ones and make sure you have enough data to calculate robust mean waveforms.

Here are the key findings 🎯 📊 :

⬆ 📈 Increased Inter-Trial Variability: the markerless system exhibited an increase in inter-trial variability of up to 22%. It seems that the markerless pose estimation pipelines introduce additional variability in kinematic data ontop of the natural variability.