Magic On The Internet – Don’t Stop Believin’

Continuing the auction theme, today we'll look at a genius study by an economist called David Lucking-Reiley. In the mid-nineties, just as online trade was really starting to take off, he analysed the sale of Magic:  the Gathering trading cards on the internet. His paper (published in AER, 1999) swiftly became a classic.

http://www.1zoom.net/big2/3/142632-stranger.jpg

Lucking-Reiley compared 4 different types of auction:

English

The most common type of auction. The bidding starts low and people make higher bids if they wish. The winner is the last surviving bidder. They win the prize and pay the winning bid.

Dutch

The starting price is set really high. It is then lowered bit by bit until someone bids. Thus the first bidder wins the prize and pays their entry price.

First Price Sealed Bid

This is not real-time. All bidders submit one bid (which is hidden from other bidders) within a set time frame. The highest bidder wins the prize and pays their bid.

Second Price Sealed Bid

Also not real-time. All bidders submit one bid (which is hidden from other bidders) within a set time frame. The highest bidder wins the prize but only pays the second highest bid.

(see last blog How To Design An Auction – Duel Of The Fates for slightly fuller descriptions).

According to the theory, the English and Second Price auctions should yield the same amount of revenue, likewise the Dutch and First Price Auctions.

http://upload.wikimedia.org/wikipedia/en/a/a7/Shadowmage_infiltrator.jpg

Lucking-Reiley initially spent $1,600 buying over 700 Magic cards. He then proceeded to sell them individually on a website devoted to trading the cards. People offered cards for sale in whatever way they liked, and a variety of different auction designs were already in use. Lucking-Reiley put up his cards for sale using the different auctions to see if they yielded the same price.

The website was already so popular (with over 20,000 messages per month) that Lucking-Reiley didn’t influence the market price. Because he didn’t tell anyone that it was an experiment he was able to accurately observe behaviour in the real world. 

The results were surprising. The English and Second Price Sealed Bid auctions were roughly the same, but the Dutch auction raised 30% more revenue on average than the First Price Sealed Bid auction.

http://mirpg.com/wp-content/uploads/2011/02/magic1.jpg

Thus if you are going to auction something, use the Dutch Auction rather than the First Price Sealed Bid. However, the experiment didn’t test to see whether the Dutch auction raised more than the English or Second Price Sealed Bid; sadly we can’t conclude which auction is best overall.

Despite this, the paper is still brilliantly innovative. David Lucking-Reiley was one of the first economists to use the web to observe how we behave. Crucially, he did so without having remove people from real life and put them in a lab (which could potentially alter behaviour). 

He also made a profit of $400. Lucky him!

Recommended listening:

Don’t Stop Believin’ by Journey