Guest post - A glimpse of the future: predicting which company will win Seedcamp

After seeing that someone had created a prediction market about Seedcamp Week winners on Hubdub.com, we asked the site’s creators if they might want to comment in a post. Here’s what they came up with –Morland

Mirror, mirror on the wall, who’s the fairest startup of all? Or rather, which of the 20 finalists will win this year’s Seedcamp and be crowned Europe’s hottest seedling? The competition is tough, but to add a bit of fun to the mix Hubdub is running a prediction market to forecast who will be taking the podium. Check it out here.

Seedcamp08: Which company will win overall best seedling?

Users bet play money on the company they want to back and win more if they are right. As well as letting users compete at predicting the future, the site combines everyone’s predictions into a forecast of who will win (as it does for thousands of other news questions like Who will be the next President of the US?, How much will Digg sell for? or Will Britney have a top ten single in 2008? But are the results of any use?

Academic research suggests they are. Prediction markets have consistently outperformed polls in election predictions because of a number of inherent advantages: people can express their confidence by the amount they wager, they output a real probability, and they can react immediately to developing news. In the same way stock prices jump when significant news breaks around a company, Hubdub’s prediction markets change over time as information emerges. This happens because people change their minds, cash out old predictions, and make new ones. For example, the market on which sport the US flag bearer in the Olympics would be from shows a nose-dive for the Swimming option when it was announced Michael Phelps wouldn’t be in Beijing for the closing ceremony. Other markets on certain events occurring by certain dates (a popular format on the site) like this one about Brad and Angelina splitting up often trend gradually towards the correct outcome as the deadline approaches, as you’d expect.

A recent volatile and unpredictable market was the one on the winner of TechCrunch50 – its lead changed several times as new companies pitched or got positive write-ups. Could we see a similar pattern for the Seedcamp market as ideas evolve over the week? Perhaps. Though this information will not be public, research has shown only a small number of people (15-20) are needed to generate accurate forecasts and there will likely be that number or more ‘insiders’ making predictions on the site – not least the companies themselves! So, if you fancy yourself as a bit of a tech pundit who knows how to pick a winning start-up and want to put your virtual money where your mouth is go make your Hubdub predictions. As for the companies themselves, there’s still everything to play for – and all of us here at Hubdub wish them the best of luck.

Hubdub.com is an Edinburgh based start-up that launched in January 2008. CEO Nigel Eccles will be at Seedcamp next week