Saturday, September 01, 2012

The Wisdom of Crowds

I have read a fascinating article, in of all places People Management (September 2012) about predictive markets. Basically the article wasn't exactly well written but the concept was strong. The article may not have been well written because it was extracted from a book, Oracles: How Prediction Markets turn employees into Visionaries, by Donald Thompson (published by Harvard Business School Publishing).

The basic concept is that a few technology and finance companies have been running internal stock markets where any employee can use fictional dollars to invest in ideas. Anyone can post an idea which is good for a couple of reasons. Firstly, there is no hierarchy in ideas, it doesn't matter at what level in the organisation you are, you can post an idea and hopefully people will listen. Secondly, because the buying and selling is anonymous just because an idea is from a senior manager doesn't mean you have to support it, unlike in most meeting where if you want to get ahead you need to support the boss (however dumb their ideas are - after all your next promotion might depend on them).

Essentially good ideas are valued commodities and bad ideas are junk stock. This approach frees up innovation and allows companies to see what ideas might work best. There is also a fun element in that if you invest wisely you will become a market leader especially if your ideas work in practice. Another part of life that has now been gameified.

Anyway as is my was I thought about this and decided I could create one of those. So I have created my own model and will build a small website for people to play with. If the model is any good it may become the next big thing or as is more usual it will be become another vaguely interesting project on the repository for all my mini projects.

Experimental Predictive Market The Rules V1.0

  • Each stock is a question i.e. "Workers in the UK should be given 1 extra bank holiday?".
  • Each stock initially lists at a price of £10 with 1000 shares for sale (Numbers would change if it ever became really popular).
  • If no one buys the stock, the price will drop at a rate of 0.04p per day (the stock and therefore idea is unpopular).
  • If someone buys stock the price rises at a rate of: 'Number of shares bought * 0.04p'.
  • If someone sells stock the price falls at a rate of: 'Number of shares sold * 0.04p'.
  • Trading limit of 100 shares per transaction.
  • If price above £10.00 stock rises at a rate of 0.04p per day (Idea is popular and will keep rising until investors tempted to sell - Price will hopefully stabilise at the value of the idea)
I will admit the above rules are a bit rough and ready and may need refining over time. Think of it as a bit of a starter for ten.

Of course I accept no responsibility if I never finish this project, I have a habit of getting distracted by real life sometimes.

No comments: