Background and purpose: Historically, sports training methodology, including soccer coaching, has been influenced by the pedagogic tend dominant in each epoch. Of late, the literature has espoused the idea that human learning is individual, nonlinear, and complex. Nonlinear Pedagogy (NLP) takes into account such nonlinearity and complex phenomena and facilitates each person's learning through a methodology of manipulating constraints. This learner-environment-centered approach conducts players and teams to a tactical behavior emerging through playing modified versions of games using a rich range of different constraints, for example the introduction of artificial rules. Despite the attractions of training based on constraints to research, few studies have empirically assessed the effects of these programs from a tactical perspective. The present study applies the NLP pedagogical principles of representativeness and constraints manipulation to provide youth sport academy coaches with empirical evidence of the training effects of modified games on team tactical behavior. Specifically, this study aimed to assess the short-term training effects of modified games introducing artificial rules on team tactical behavior during the team possession game phase in two youth soccer categories. Methods: Thirty-two Spanish male youth soccer players from the under-14 (U14: n = 16; age: 13.0 ± 0.4 years; playing experience: 6 ± 1 years; height: 1.56 ± 0.07 m; body mass: 46.7 ± 5.9 kg) and the under-16 (U16: n = 16; age: 15.9 ± 0.4 years; playing experience: 8 ± 1 years; height: 1.70 ± 0.07 m; body mass: 57.9 ± 8.3 kg) teams of the same club participated in the study. The natural groups were not modified for the study and the participants had some prior experience with modified games. Each team was divided into two groups who played each other during a short-term intervention distributed across four training sessions: control (CON: free play without restrictions) and experimental (EXP: modified games introducing artificial rules). Team tactical behavior was assessed, including the goalkeeper (Gk) and during the team possession game phase, using a local positioning system (LPS) before and after the intervention in an 8-a-side game with the official offside rule. Findings: Linear team tactical behavior change in centroid position, dyads, and total area values barely changed (d = trivial -- small [0.01-0.34]) after a short-term intervention of free play and modified games. Team tactical behavior tended toward greater regularity (sample entropy decrease) after artificial rules were introduced in comparison with free play without restrictions. Youth sport academy coaches who opt for organized and regular behavior during the team possession game phase could employ modified games introducing artificial rules to build up repeatable tactical patterns. Conversely, those who prefer being more chaotic and unpredictable in the game could take advantage of free play without restrictions.