This paper presents an exploratory study of two contrasting cases observed longitudinally, using the different versions of the LTP procedure from infancy to adolescence, to evaluate the continuity of family processes, coparenting, and children’s triangular communication. Two cases, matched for SES and gender of the children, were selected from a sample of 50 volunteer families. One family had a functional alliance, while the other had a detouring coalition. We begin with descriptions of the family interactions, the coparenting alliance and the children’s triangular communication at 18 months, followed by brief descriptions of their LTPs at 3 and 9 months as well as at 5 years. Finally, we present the LTP for adolescents and their parents, which involves a discussion task rather than play (LTP/Discussion). Detailed excerpts contrast the interactions of the two families once the children were 15 years old. These descriptions show the continuity of family processes, particularly “three-way” communication in the family with a functional alliance vs. “two-against one” communication in the family with a coalition. Data from microanalysis of gaze and affect, during the LTP/Discussion, provide a second way of assessing communication patterns. The microanalysis revealed great discrepancies between the families at adolescence, not only in terms of the pattern of gaze interaction, but also with respect to the valence of affects. In the functional alliance, fun and affection dominated and gaze patterns were typical of three-way communication. Gaze patterns of restricted communication, lack of empathy and hostility characterised the detouring coalition.