Fall detection has been widely studied in the literature; the main part of the studies are focused on people suffering some severe illnesses. As a consequence, the majority of the solutions relies on sensors located in positions where the ergonomic issues are in compromise. This research represents a review of a published work concerning fall detection on healthy people using an accelerometer located on a wrist, which is more comfortable and usual gadget location. Using event detection and feature extraction, a set of variables are generated. These variables, together with on-class classifiers, conforms an anomaly detection system for fall detection. The experimentation has considered two publicly available data sets and the obtained results are highly competitive. The present contribution represents a keynote about our published research.