Collective intelligence applied to equity investments
I was interested in the direct use of “collective intelligence” for investment. It turns out that if you organize and summarize the opinions of many investors in a special way, you can get investment forecasts that are superior in their effectiveness to the results of the best professional analysts.
For the curious, this is how the page with stock forecasts turned out, according to the collective mind at the moment. And the details are under the cut.
In recent years, many concrete examples of success have been accumulated in predicting a variety of facts based on the theory of “collective mind” or “wisdom of crowd”. At Google, for example, employees make predictions about events such as the launch dates of new projects and offices.
On Habré "wisdom of crowd" in various applications was already discussed. For those who have not read these discussions, a brief summary - the theory states that, under certain conditions, a group of people can predict events better than any individual expert.
Like many Russian investors, I watched the quotes of Russian stocks in recent times in tension. Our market has lost two-thirds of its capitalization in just three months, despite encouraging statements from businessmen and politicians. Moreover, if you restore the picture of what happened bit by bit, it is clear that some experts made a fairly accurate diagnosis of what was happening and foresaw many steps of the stock market. But their voice was drowning in the general informational noise. In order to make better stock forecasts, you need to find the most reasonable voices in the crowd rumble. Briefly, the idea is as follows.
People who have an idea of investing in Russian stocks share their forecasts with the community in a short form like “Gazprom will be better than others in a few months.” The goal of the site is to create a transparent mechanism for identifying the best investors who regularly make correct forecasts.
As an added bonus, you can create stock forecasts based on a generalized view of the market for the entire community - a “collective mind” in action. As a result, it turns out that each member of the community can make a personal investment decision by looking at the forecasts of the most successful investors, as well as a generalized rating of shares.
With a mechanism for selecting the best forecasts, it was relatively simple. Since we have already arranged display of data on stocks from the RTS and MICEX exchanges in TIKR.ru, it only remained to create a process for collecting and comparing forecasts with data on real quotes. The only less trivial question is how to compare stock prices?
In a good time for the market, the market index (a basket of the most significant stocks) goes up. It's easy to make money these years, but it's hard to be better than the index. In such a bad time, as now, the index goes down, but not everyone can lose less than the index, not to mention profit. Anyone can achieve "average" profitability by investing in stocks repeating the structure of the index, or in the index mutual fund. Incidentally, in this sense I am amazed by the advertising of our mutual funds - they rarely compare their indicators with the corresponding indices. As a result, to compare the success of forecasts in our project, the main MICEX index was chosen as the most representative index of the Russian stock market.
With the mechanism for selecting the best investors, it turned out to be a little more complicated. Initially, I wanted to make a mechanism similar to karma from Habr. But in investment, your karma, in fact, does not entirely depend on your ability to make sound judgments about investments and the community's votes for / against your statements. Much more important in investments is the ability to make a forecast on which you can earn more money than when investing in the index.
Here I turned to the analysis of similar projects in the west. By the way, I was especially impressed by the approach of the Microsoft investment community, CAPS MSN, created as a joint project with caps.fool.com. But at the same time, I did not want to become a training simulator or an exchange game, because we already have a serious audience of thousands of real investors in Russian stocks.
Following the example of the sites listed, it was decided to measure the success of the investor in the account and accuracy. Roughly speaking, the account for each forecast is equal to the difference between the stock yield and the change in the MICEX index; for each investor, the forecast results are aggregated. Accuracy is the proportion of correct forecasts of the total number of investor forecasts. As a result, we get the rating of investors, and each member of the community sees how much better or worse his forecast is relative to the forecasts of other participants.
For each stock, all forecasts for this stock are summed up, taking into account the investor rating. It is clear that the better the investor generally invests, the more influence he or she has on the rating of the stock. The received rank of the action reflects the community's collectively reasonable forecast for each given moment. The direct benefit of this approach is that the success of investment forecasting becomes transparent for all members of the community, and each investor can find new investment ideas for himself, relying on collective wisdom.
It should be noted that amateurs from Habr who know only one industry or company well can try their hand at forecasting. I note in advance that only those users who have shared their forecasts with other community members have access to unique summarized results on stock forecasts. But it is possible to compete in the ability to predict with professionals - TIKR.ru contains forecasts of analysts from large investment companies, such as Unicredit Aton, Otkritie FC, BCS and others.
You can talk about the intricacies of our implementation for a long time. But in this place I would like to invite all interested hub users to a joint discussion, because the exchanges will close at 18:45 and if the forecast is made after this time, then we will have to wait for tomorrow morning to get the first real rating.
Most of all, as proposals are usually interested in - what can be improved, done differently? Well, bugs, too, may have remained. I apologize in advance and promise to edit and answer questions in real time.
For the curious, this is how the page with stock forecasts turned out, according to the collective mind at the moment. And the details are under the cut.
In recent years, many concrete examples of success have been accumulated in predicting a variety of facts based on the theory of “collective mind” or “wisdom of crowd”. At Google, for example, employees make predictions about events such as the launch dates of new projects and offices.
On Habré "wisdom of crowd" in various applications was already discussed. For those who have not read these discussions, a brief summary - the theory states that, under certain conditions, a group of people can predict events better than any individual expert.
Like many Russian investors, I watched the quotes of Russian stocks in recent times in tension. Our market has lost two-thirds of its capitalization in just three months, despite encouraging statements from businessmen and politicians. Moreover, if you restore the picture of what happened bit by bit, it is clear that some experts made a fairly accurate diagnosis of what was happening and foresaw many steps of the stock market. But their voice was drowning in the general informational noise. In order to make better stock forecasts, you need to find the most reasonable voices in the crowd rumble. Briefly, the idea is as follows.
People who have an idea of investing in Russian stocks share their forecasts with the community in a short form like “Gazprom will be better than others in a few months.” The goal of the site is to create a transparent mechanism for identifying the best investors who regularly make correct forecasts.
As an added bonus, you can create stock forecasts based on a generalized view of the market for the entire community - a “collective mind” in action. As a result, it turns out that each member of the community can make a personal investment decision by looking at the forecasts of the most successful investors, as well as a generalized rating of shares.
With a mechanism for selecting the best forecasts, it was relatively simple. Since we have already arranged display of data on stocks from the RTS and MICEX exchanges in TIKR.ru, it only remained to create a process for collecting and comparing forecasts with data on real quotes. The only less trivial question is how to compare stock prices?
In a good time for the market, the market index (a basket of the most significant stocks) goes up. It's easy to make money these years, but it's hard to be better than the index. In such a bad time, as now, the index goes down, but not everyone can lose less than the index, not to mention profit. Anyone can achieve "average" profitability by investing in stocks repeating the structure of the index, or in the index mutual fund. Incidentally, in this sense I am amazed by the advertising of our mutual funds - they rarely compare their indicators with the corresponding indices. As a result, to compare the success of forecasts in our project, the main MICEX index was chosen as the most representative index of the Russian stock market.
With the mechanism for selecting the best investors, it turned out to be a little more complicated. Initially, I wanted to make a mechanism similar to karma from Habr. But in investment, your karma, in fact, does not entirely depend on your ability to make sound judgments about investments and the community's votes for / against your statements. Much more important in investments is the ability to make a forecast on which you can earn more money than when investing in the index.
Here I turned to the analysis of similar projects in the west. By the way, I was especially impressed by the approach of the Microsoft investment community, CAPS MSN, created as a joint project with caps.fool.com. But at the same time, I did not want to become a training simulator or an exchange game, because we already have a serious audience of thousands of real investors in Russian stocks.
Following the example of the sites listed, it was decided to measure the success of the investor in the account and accuracy. Roughly speaking, the account for each forecast is equal to the difference between the stock yield and the change in the MICEX index; for each investor, the forecast results are aggregated. Accuracy is the proportion of correct forecasts of the total number of investor forecasts. As a result, we get the rating of investors, and each member of the community sees how much better or worse his forecast is relative to the forecasts of other participants.
For each stock, all forecasts for this stock are summed up, taking into account the investor rating. It is clear that the better the investor generally invests, the more influence he or she has on the rating of the stock. The received rank of the action reflects the community's collectively reasonable forecast for each given moment. The direct benefit of this approach is that the success of investment forecasting becomes transparent for all members of the community, and each investor can find new investment ideas for himself, relying on collective wisdom.
It should be noted that amateurs from Habr who know only one industry or company well can try their hand at forecasting. I note in advance that only those users who have shared their forecasts with other community members have access to unique summarized results on stock forecasts. But it is possible to compete in the ability to predict with professionals - TIKR.ru contains forecasts of analysts from large investment companies, such as Unicredit Aton, Otkritie FC, BCS and others.
You can talk about the intricacies of our implementation for a long time. But in this place I would like to invite all interested hub users to a joint discussion, because the exchanges will close at 18:45 and if the forecast is made after this time, then we will have to wait for tomorrow morning to get the first real rating.
Most of all, as proposals are usually interested in - what can be improved, done differently? Well, bugs, too, may have remained. I apologize in advance and promise to edit and answer questions in real time.