Forecasting Auction Outcomes: Assessing the Efficacy of Statistical Models in Sealed-Bid Auctions

This thesis investigates the predictive accuracy of several statistical models have in sealed-bid auctions under different market conditions. This paper studies all such models through comprehensive literature review and empirical study, including Friedman’s model, Gates’ model, Skitmore’s multivariate model, Hanssmann and Rivett’s Model, Knode and Swanson’s stochastic dynamic programming model, option pricing approach, and basic mean models (i.e., arithmetic, geometric, harmonic). The adopted research method relied on computer simulation to develop synthetic datasets representing real auctions settings. This has in turn been able to give a comparative evaluation of the effectiveness of each model. The results show that predictive accuracy depends on the model considered, the number of bidders, and the variability of bids. The study highlights the conditions under which each model can be optimally used to provide valuable insights for auction organizers and bidders seeking to optimize strategies in competitive bidding scenarios. This finding adds to the theoretical discussion of auction strategies and also finds practical implications for auctioneers to improve bidder success rates through the use of data-driven decision-making tools. This research not only bridges the gap between theoretical auction models and real-world applications of these models, but also demonstrates the importance of integrating advanced statistical techniques with information technology.