AGORA. For decision making.

Mark Roulston, Senior Data Scientist •

Blog agor2
Imagine what value the dispersed human knowledge and expertise found in your organisation could provide if it were aggregated into a structured forecast. Our new prediction tool, AGORA, can do just that.

Harry. S. Truman reputedly asked for a one-handed economist, because he was fed up with economists hedging their advice by saying, “On the one hand…” and “on the other hand…”.

John F. Kennedy quoted George Bernard Shaw saying, “If all the economists in the world were laid end-to-end, they still would not reach a conclusion.”

Both presidents were commenting on the problem of extracting forecasts from economists and combining them to create reliable information they could act upon. Because, no matter how well-honed your decision-making skills, the decisions you make are only as good as the knowledge that informs them.

This can be a real challenge when the knowledge and expertise relevant to your question is widely distributed throughout your organisation.

Using the collective intelligence revealed by a ‘market’ One of the oldest known mechanisms for aggregating this kind of distributed information is through a ‘market’.

In his essay “The Use of Knowledge in Society”, published in 1945, economist Friedrich Hayek described how the knowledge of many individuals is summarised by market prices. Because whilst markets exist to allow people to buy and sell things such as potatoes, companies or debts, their ability to reveal information is a very useful side effect; which Hayek described as a “marvel”.

More than thirty years later, the proliferation of computers allowed economists to create markets purely to obtain information: these markets became known as ‘prediction markets’.

Prediction markets - a structured way to aggregate distributed information In a prediction market, contracts that pay out 1.00 unit if a specified outcome occurs, are traded. If the prevailing price for an outcome is 0.40, it implies a 40% chance of it occurring.

If you think the true probability is more than 40% you buy the contract, pushing the price up. If you think the probability is less than 40% you sell the contract, pushing the price down. Through the individual actions of each participant, a market price emerges which reflects the consensus view of the probability.

The modern-day use of prediction markets Prediction markets have been used by companies wanting to collect information dispersed among their employees. Hewlett Packard (HP) was one of the earliest experimenters, using internal markets to forecast sales. Lew Platt, the CEO of HP, summed up the problem of capturing distributed knowledge when he said, “If only HP knew what HP knows”.

Google has used prediction markets to predict project timelines; the pharmaceutical company Eli Lilly used them to predict enrollment in drug trials, and Taiwan’s Centers for Disease Control tested a prediction market of medical professionals to forecast epidemics.

Introducing AGORA, an easy-to-use prediction market tool from Hivemind Hivemind has now created AGORA, an online tool that creates private prediction markets, where useful insights can be shared, collected and synthesised into a consensus.

AGORA can be used to create simple binary markets asking if an event will happen or not, and it also supports markets that generate probability distributions for numerical quantities, e.g. sales forecasts.

There is information and tacit knowledge dispersed throughout most organisations. Prediction markets on AGORA can be used to capture this information and turn it into business intelligence.