| Abstract: | In this paper, we show some counter-intuitive observations regarding the open-loop policies in a Bayesian dynamic pricing problem. Specifically, while the common intuition that a policy incorporating more information performs better continues to hold under ample inventory for sale, it breaks down in the case where inventory is limited. This can be explained by the unique feature of selling a limited amount of inventory: information updating may stop prematurely when inventory is depleted, rendering certain demand information being inconsequential. |
