In the initial sorting step, it is important to retain as many accepted contenders as possible. However, the accepted contenders sorting first gives no useful information to the caller. Re-sorting by reward rank ensures the best contender is first. In the Processor, we will use this to then accept the highest ranking contender (even if it was not the local stake winner).

An alternative approach would be to return the best contender ID in addition to pollable contenders list, but this API is less intuitive and the implementation gives no benefits in simplicity.