illuminating science

Well, in the end I said pretty much all I wanted to say on auctions and Ebay. The whole reason I first posted about Ebay was because of an item I was bidding on, and some interesting bidding psychology. It retails for $80 here, but goes for just $10 (!) on Ebay - or, at least, that’s the starting price, and it goes up from there. I saw that a few of them had gone for between $20-30, so I ended up bidding $31.73 as my top bid - after all, that’s still half price. I missed out on two of them due to snipers (people who put in bids at the very last second, giving you no chance to respond) - very frustrating! Of course, logic dictates that if I put in my maximum bid, then I can’t be disappointed if I miss out. But on the other hand, for just a dollar more, I could have the item - and what’s a dollar, compared to 30?

In the end, I put in a bid for $37.73 (again, note the psychology of my maximum bid!) and was beaten again - only to have the highest bidder pull out, and for them to offer me the item for $37.73. I ummed and ahhed, and eventually decided to take it - after all, I had been prepared to pay that before, and since all the others had gone for at least $30, it wasn’t a bad deal.

Now that I’ve bought it, of course, suddenly demand has died off - they’re going for $10!!! $10 - can you believe it?! Overpaid by $30…GAH! There’s probably a business opportunity here… Anyway, if I’d followed my normal practice of studying the market very thoroughly first, instead of jumping in, I would have gotten a much better deal - but I was worried that this guy would only have a few of them to sell at that price. Instead, he seems to be churning them out at 2 a day - don’t know how he makes any money!

But it made me wonder - how can both the auctioneer and the bidder ensure that they get good prices for their items? A little bit of digging came up with the Vickrey auction. Everyone makes secret, sealed bids - then the winning bidder pays the amount bid by the second highest bidder! (The second highest bidder neither pays nor receives anything - they simply set the selling point.) It’s actually quite easy to prove that the optimal strategy in this auction is to bid what you think the item is worth - any more, and you pay too much, any less and you don’t gain anything but might miss the item. Neat, hey? Now, think about whether the auction would work if you didn’t have silent bids, and ran it like a normal auction (going once, going twice…) except with the winner paying the second highest bid. Could you beat the system?