Similar to the design-build delivery system, the alliance delivery system is used here to denote a special form of partnership between the owner and the design-build team, demonstrated through the substantial involvement of the owner. Design-Build/Alliance delivery systems have been implemented successfully for private projects. However, their use in public work has been hindered by many issues including the presence of qualitative selection criteria and the absence of a scientific evaluation process to select the winning team. This paper lays out a framework which facilitates selecting the best alliance team for a project. Monte Carlo simulations were used to vary all the factors relevant to the decision problem and to uncover biases present in the evaluation. These may arise due to any unfairness of the model represented by the weights of the criteria or any possible discrimination from the evaluators. Use of this tool would complement the high performance of design-build/alliance project delivery systems and further enhance it by ensuring that the selected team is the best one for the job, thus lowering project risks. A case study dealing with a large utilityproject illustrates the developed framework. Copyright ASCE 2009.