It turned out that Shiau's predictions weren't perfect--on average he was about 15 percent off on the sales of Nintendo's DS and Wii, Sony's PlayStation 3 and PlayStation Portable and Microsoft's Xbox 360. But he also wasn't that far off predictions from Wedbush Morgan's Michael Pachter, one of the most quoted industry analysts, who himself had been about 10.6 percent off sales on the same consoles.
And a month earlier, Shiau's predictions had actually been better than Pachter's. Shiau had been off by about 15.8 percent, while Pachter had missed by 34.9 percent.
Actually, Shiau's predictions weren't his own. They were those of the more than 4,000 members of SimExchange, the prediction market that Shiau started in November that is likely the first such market devoted entirely to the video game industry.
Ais a sort of fantasy stock market that is increasingly being used to predict everything from internal sales numbers at Fortune 500 companies to election performances by politicians. They are a manifestation of New Yorker staff writer James Surowiecki's concept of "the wisdom of crowds," in which large numbers of people tend to be better able to assess the outcome of something than individuals.
The SimExchange functions as a stock market, with members trading on the future sales of everything from hardware like the Xbox and PlayStation 3 to software like God of War and Super Paper Mario.
"Prediction market technology hadn't been applied to video games," said Shiau. "For the consumer side, there's just so many games out there, and so many reviews to read, and what I believed was that not only could a prediction market help consumers identify what games and accessories would sell the best, (it could) also help professionals."
SimExchange members don't actually spend real money on their purchases of the "stocks" (though participants in some other prediction markets do put real money on the line) and therefore they can neither win nor lose money based on the outcome. Instead, they bet what are known as "DKP," a faux currency that translates to 10,000 units sold of whatever game, console or accessory the member is betting on.
The idea is that by buying and selling DKP, SimExchange members will be able to, in aggregate, predict how well the products they're betting on do in real life. It may be an odd concept, but it has been proven in other prediction markets to be fairly accurate, so long as the participants are betting something that matters to them.
And while SimExchange members don't bet real money, they do accumulate ratings based on how accurate their bets are. And because members are predisposed to be interested in video games and video game mechanisms, a competitive score may well be enough to motivate more accurate predictions, suggested Shiau.
"People are (expressing) opinions by putting a resource on the line," he said "They only make them strongly when they put something at risk. People care about accumulating the virtual currency. They're competing in global rankings, or with their friends."
In the exchange's best results from its predictions for April, members had bet that sales of the PlayStation Portable would come out at around 190,400 units. In fact, NPD reported Thursday, Sony sold 183,000 PSPs. So SimExchange members were off by just over 4 percent. But by comparison, Pachter had predicted sales of 200,000 PSPs, a miss of 9.29 percent.
"Brian Shiau seems to have created a nascent and vibrant community," said Chris Masse, the author of Midas Oracle, a blog about prediction markets. "(There's) a strong emphasis on using the SimExchange as an educational tool to teach youngsters about the stock and futures mechanisms. I like this a lot."
Shiau said SimExchange is built to provide members with as much information as possible. Each product that is up for trading has a channel on the service's site where members can discuss the product and post news articles, screenshots and any other relevant information. That way, each member can have access to the same information and have the best chance to bet accurately on the product's performance.
And while SimExchange is free to play, its business model is to sell video game-based ads.
The SimExchange is currently aimed at casual enthusiasts, but Shiau hopes that its accuracy will improve as more members join and that it could someday provide a valuable resource to the industry itself.
He acknowledges that there are some things standing in the way of that goal, however.
"The biggest obstacle is that most people don't know what a prediction market is," said Shiau. "So it's very difficult to pitch it to someone without them understanding how a stock market (provides useful information)."