The objective of this study was to design, validate and update an observation tool to analyse the technical-tactical actions by which taekwondo players win points. An ad hoc observational tool was developed for subsequent use in HOISAN software by viewing seven finals (14 viewings) in the Rome 2019 Grand Prix, collecting data (1,382 technical/tactical actions) from both winners and losers (women, n = 3; men, n = 4). An observational methodology based on a nomothetic, follow-up and multidimensional (N/F/M) observational design was used. In the statistical analysis, to check the validity of the generalizability analysis, the Category/Observer (C/O) and Observer/Category (O/C) models were employed, and to determine reliability between observations (intra- and interobserver), the Pearson, Spearman, Kendall's tau-b and Cohen's kappa correlation coefficients were applied. One point were awarded for every penalty given against the opponent. Two points were obtained for a circular technique to the trunk protector (in winners) or by scoring a point for a punch plus a penalty against the opponent, three points in melee actions, anticipatory actions with the left side (winners), or after a punch action, and circular technique (losers). Four and five points were only obtained by winners, in a direct attack with the right leg, turning (five points) or simultaneous (four points). The results of this study provide information on the most effective technical-tactical actions in taekwondo for scoring points in a contest.