This may be both a stats and economics question.
Let's say that I am bidding in an auction for emails. There are 10,000 + plus emails and there are five bidders and each of us bids on a SINGLE email throughout the day as they become open to bids. The highest bidder wins the auction. Our revenue comes from us turning around and selling that email to someone else a few seconds after buying. However, the highest bidder could turn down a email after he has won if he can't turn around and sell it. So I'm trying to calculate expected revenue for this complex auction system. Initially, I just took the average revenue of the past 30 email that we won and sold. However, this ovrt-estimates our revenue because it doesn't account for email we won but ended up declining. I also was not considering that expected revenue should include our bid on the emails within the auction.
What would be a good strategy for calculating expected revenue in such a system?
I realize that trying to make statistical sense of auction systems can often be a futile task. However, I think this strategy should work better than just having some random bidding strategy.