Abstract: This paper develops a player evaluation framework that stresses the importance of accounting for complementarities between teammates when evaluating players. This is done by developing a probabilistic model of a basketball possession as a progression of events, where the probability of each event’s occurrence is determined by the offensive players’ skills, the defensive players’ skills, and the complementarities between the skills of teammates. Evaluating players using this frame- work allows me to assess the substitutability between different game actions, the lineup-specific value a player brings to a team, and the players that are the best and worst teammates. It also al- lows me to separately identify the individual effect from the effect teammates have on a player’s statistical production, and to evaluate whether player complementarities are valued in the market for NBA players in terms of higher salaries. I find that complementarities are under-valued, and that players are instead paid mainly for their individual statistical production.