Crowd Against Analysts

The developer of Cindicator app holds the opinion that the ‘crowd indicator’ generated by the same will soon replace standard recommendations provided by analysts. RBC looked into the matter to decide whether the app would be useful for private investors.

Cindicator (the app has been offered in AppStore since 10 May) is a platform for the users to share their opinions related to the outcomes of various events. E.g., they’ll decide on whether the S&P 500 Index drops below 2050 level before 23 July or the new Ghostbusters film earns over $50 mln the weekend after its first release.

The app comes with the analytic platform to be available for customers (e.g. hedge funds, analytic agencies or private investors) in late August. They will be able to get access to the so-called Crowd Indicator, says Mikhail Brusov, CEO of the Project. “The indicator means the weighted average probability of the event occurrence (on various scenarios) provided by all users of the app,” he explains. “Moreover, if people keep answering the questions for an extended period, the indicator will be even more accurate than the opinion of any given expert,” says Brusov. He referred to the book Superforecasting: The Art and Science of Prediction by Philip E. Tetlock and Dan Gardner, which he’d discovered when working on the project at Starta Accelerator in NYC.

The book was based on a two-year experiment in the course of which questions were asked to several thousand respondents related to various social groups and possessing various levels of knowledge. The experiment prescribed that no participants could get to know the answers given by other respondents, continues Brusov. The questions related to different fields, namely: sports, politics, economy. As a result, according to Brusov, they managed to prove the idea of the so-called wisdom of crowd effect — in 80% of cases, the average weighted opinion of respondents was more accurate than the answer of a single most accurate expert in that professional field. Moreover, about 2% of respondents turned out to be superforecasters, i.e. people who tend to make more accurate predictions than the rest of the group, explains Brusov.

How the app works

The app looks like news feed specifying the events with the outcomes to be made public in the nearest future, i.e. ‘agenda for tomorrow’ (as Brusov says). The description of each event contains a question, which can be answered only with a “yes” or “no.” For example, “Is the price of Mariott International Inc. shares going to rise to $68 before 12 July?” — “Yes, it is,” “No, it isn’t.”

Besides, the user estimates the probability of the event occurring in percent. In Cindicator, the initial probability value is set at 50%, i.e. the probability of an event occurring is supposed to be equal to 50%. If the respondent strongly believes that such probability is high, e.g. the government is going to ban using Tesla’s autopilot system, after having investigated the fatal crash involving the driver, he or she sets the value above 50%. If the user thinks that the event is unlikely to happen, he or she sets the value below 50%.

Brusov recommends the users to focus on the links contained in each question. For example, the link given in the question about the next national coach of UK is a reference to the interview with the prospective candidate, and the one related to the question about the cost of shares of any company — to the data about their cost fluctuation over the past year.

Each user is assigned a rank depending on the accuracy of his or her predictions. In case the respondent had answered the question “Will Andy Murray become Wimbledon champion?” by setting the probability value at 87%, he or she would have gained 37 points (for Murray became a champion, indeed). If he or she had set the probability value at 30% (meaning that in his or her opinion Murray is unlikely to become a champion), he or she would have lost 20 points. At the end of the month, 2% top rated users (superforecasters) will be given money for their predictions. At the moment, the prize pool is a little bit more than $1K (the company budget allocated for marketing). We are planning to increase this amount in the future — a portion of funds obtained from customers for their access to Crowd Indicator will be used to build-up the pool, says Brusov. According to his estimates, the pool will be equal to approx. $5K by year-end.

Starting from August, investors will be able to use Crowd Indicator generated by the app. For $200 per month, you can have the access to forecasts related to the current events and also all data concerning previous events. The users having paid the corporate fee (the fee amount still remains unknown) will be provided with the opportunity to add events to their news feed and also obtain the results only from superforecasters. Brusov is sure that the project will be focused mainly on US hedge funds, as they need new sources of information. They (hedge funds) may be specifically interested in the opinion of a crowd of people unrelated to each other, thinks Brusov. “At the moment, traders rely on the forecasts provided by financial analysts. But they are made with due regard to one and the same information, being in the public domain. Crowd Indicator, in its turn, can be used as an extra prompt for decision-making, he says.

In June, app users provided most accurate forecasts concerning 115 situations and made a mistake concerning 15 cases; Brusov gives arguments in favor of the crowd indicator efficiency. In 70% of cases, the indicator value was a well-defined figure, i.e. above 60% or below 40%, he adds. According to his version, investors tend to take into account such clear signals.

Brusov says that in the future Crowd Indicator will pay attention not only to predictions made by users with the help of the app but also to their messages in social networks. Approximately a half of their messages reflects their opinion on the domestic or global event, he explains. Such information may be of great value for markets, if it is treated in the right way, tries to convince Brusov.

Next week, the app originator is planning to run a demo merchant account and use there only the signals obtained from the platform. “The demo account will be open to anyone and any person will be able to see the balance and all transactions made using the account. Moreover, we are going to provide comments concerning each transaction and explain our reasons for making the deal, says Brusov.

Should One Rely on Wisdom of the Crowd?

“The crowd can know exactly the outcomes of any event under no circumstances. In this situation, it is more correct to speak about the attitude of the crowd rather than its wisdom, as it is the attitude that is measured”

shares his detailed opinion Andrey Rossokhin, Head of Psychoanalysis and Business Consulting Department at HSE. The expert also points out that the crowd may be controlled.

Sergey Troshin, IT Director at Exante Investment Company, has the same opinion.

“The most difficult thing is to understand whether anyone tried to change the public opinion with a goal in his or her mind or such an opinion has been shaped with due regard to the information which is more or less unbiased.”

Andrey Kochetkov, the analyst at Otkrytie Broker, considers the app samples as unrepresentative. To his mind, poll results may differ from real opinion of the majority. To obtain representative samples, only in Russia, we’ll need the opinion of several million people, believes Kochetkov. However, as Brusov said, the app is currently being used by 2 thousand people from different countries. Forecasters’ qualification also invites experts to ask some questions. Only a team of qualified professionals can be successful in forecasting the outcomes of an event, Rossokhin from HSE is rigid in his beliefs. “It would be of greater interest if the system made an estimate taking into account the opinion of great many professional market participants,” agrees Kochetkov (Otkrytie Broker).

However, Rossokhin agrees that Cindicator is a useful tool which can be applied by financial experts. And Troshin from Exante assumes that Crowd Indicator may become a tool to develop the investment portfolio in the long term. “But here we speak exactly about the portfolio and not about investing all the funds in one market segment,” he emphasizes.

Original article: http://www.rbc.ru/newspaper/2016/07/15/5786621f9a7947c55671896c