Consider the auctioneer’s point of view. It sees that the bidders are bidding their true values V1,V2, …, Vn. The bidder with the highest value will win such an auction and the price will be whatever the second-highest value is. This is where the order statistics become useful again. The winning bidder’s value is V1:n and the second-highest value is V2:n. So, the expected price (i.e., seller’s revenue) is E[V2:n]. E.g., if there are 4 bidders and the bidders’ values are i.i.d. uniform over [0,1], then the expected revenue in the auction will be E[V2:n] = 0.6. (See the earlier discussion of order statistics for verifying this result.)