The algorithm that foresees stock market trends

Anonim

Playing on the stock exchange is a rather risky activity, strongly advised against those who cannot afford to lose money. But is it possible to attempt a scientific approach to the market that minimizes risks and maximizes earning opportunities?

So far, many have tried, but the perfect method to "beat the dealer" has not yet been developed (and if someone tries to convince you otherwise, don't believe them: he is a scammer).

A team of researchers from the Tippie College of Business (Iowa, United States), however, has recently presented a newly conceived statistical model that allows forecasting the trend of the securities, in terms of upward or downward, with a more accurate margin of accuracy than at a random 50-50. The only drawback: the time window for the application of the model is only 30 minutes.


The scheme is there but cannot be seen

Michael Rechenthin and his collaborators analyzed the fluctuations of the S & P500 index throughout 2005. It is a basket containing all the 500 Standard & Poor shares and is considered representative of the entire American market. And with over 90, 000 transactions a day it is one of the most treated and offers analysts a large amount of data.
The researchers did not find any recurring pattern until the stock price fluctuates between the bid and ask values, ie those to which the traders are willing to sell and those the market is willing to pay. But as soon as this barrier is broken, the actions show a recursive and, to a certain extent, predictable behavior.
Matter of seconds
Analysts have taken the index prices at intervals of 1, 3, 5, 10, 20 seconds and 1.5 and 30 minutes after the boundaries of the spread were exceeded and found that its fluctuations are linked to the trend of the last transactions. For example, if the last transactions are two rises followed by a drop there is a 52% probability that the stock will fall in the next 5 seconds. Within 20 seconds the possibility that it drops is reduced to 43%.
An almost perfect market
The time intervals considered by the researchers are so short that they consider the trend of the stock immune from external market influences, for example information or socio-political events. But wait before you place yourself in front of the computer to trade online: the same authors of the study recall that the model has yet to be tested and validated on other securities and other markets.
You might be interested
Bitcoins and the future of the economy (we explained them already in 2011)
The tobacco economy