The paper presents a theoretical analysis of power distribution in exchange networks, supported by laboratory experiments and computer simulations. It compares two theoretical approaches: point centrality in graph theory and power-dependence principles. Centrality measures are easier to apply to complex networks but fail to predict power distribution in the studied networks. Power-dependence theory, though more complex, generates hypotheses confirmed in experiments and simulations. The study highlights the distinction between positive and negative connections in networks and the role of position in determining power. It argues that centrality measures may predict power in some networks but not others, and that power-dependence theory offers a more general framework. The paper also discusses the importance of network structure in determining power dynamics, with experiments showing that power is concentrated in central positions. The results support the idea that power is a function of position and that power use increases over time. The study concludes that power-dependence theory and structural centrality can be integrated to better understand power distribution in exchange networks.The paper presents a theoretical analysis of power distribution in exchange networks, supported by laboratory experiments and computer simulations. It compares two theoretical approaches: point centrality in graph theory and power-dependence principles. Centrality measures are easier to apply to complex networks but fail to predict power distribution in the studied networks. Power-dependence theory, though more complex, generates hypotheses confirmed in experiments and simulations. The study highlights the distinction between positive and negative connections in networks and the role of position in determining power. It argues that centrality measures may predict power in some networks but not others, and that power-dependence theory offers a more general framework. The paper also discusses the importance of network structure in determining power dynamics, with experiments showing that power is concentrated in central positions. The results support the idea that power is a function of position and that power use increases over time. The study concludes that power-dependence theory and structural centrality can be integrated to better understand power distribution in exchange networks.