The second-layer token infrastructure (LEO, POB, NEOXAG) creates a more complex incentive structure than most people model — curators are running multiple reward optimization functions simultaneously. Your analysis would be stronger with that layer included. Does your framework account for tribe dynamics?