Can we forecast the stock market?


I think this is the most common question asked by new users of the software: "can your program forecast/outperform the stock market consistently, with some significant accuracy, let say, 60-70%?". I always feel myself uncomfortable when facing this question. This feeling comes not because the program does not do something of the kind. The reason is totally different. It is in details of the question. I prefer the straightforward answer. YES, the software is able to forecast the market. But, if you are asking me whether my software can ALWAYS forecast the stock market significantly, the answer is NO. Moreover, I do not believe that you can find such a thing somewhere. If you are looking for something like this, please forget my software and this website - I can not help you here. 

Game dreamers versus skeptics

The human beings always search for the eternal truth. This is natural and obvious. If you would find the hidden key of the market behavior, you are the winner. Practically all doors of this word will be opened for you. The idea of  the ``philosopher stone`` in the stock market is irresistible; it is appealing for any representative of the human kind. I think all newcomers in trading considered different courses and programs that claimed the ability to reveal the Order in the Chaos of stock market's behavior and definitely spent some money on them.

This is one side of the picture. 

Let us look at it from another point of view - from the view of professional traders/trading  robot developers. I have never seen more skeptics than them; the cold wind of the stock market reality does not leave room for any illusion. This is not the job requirement only, this is the state of the mind - "find the way to make profit or die". Try to speak to these people about some course on hidden market patterns and add that this pattern allows to win this game with 80% probability; you figure out immediately what I am talking about. For these people, more close statement is "the opportunity is when stupid people have a lot of stupid money".

These are two absolutely different groups of users I deal with; and in this article I will try to build the bridge between them.

Efficient Market Theory - the Whole Universe in One Digit

The first rule of the market game that you should accept is: there are no everlasting rules in the stock market (I mean the rules in ways of making money with the stock market). Market modeling, market forecasting, a conscious trading is a game, and the rules of this game constantly change. Of course, there is a special science there that covers this rule. It is called Efficient Market Theory. It states that the price itself takes into account all events that happen in our World.  The price is a kind of the portrait of the whole Universe represented in one constantly changing digit.

If something happens in the world, the price practically immediately reflects this event - the stock market is efficient. In other words, if somewhere in the word some guy somehow finds a tradable hidden market pattern and starts to trade using it, the stock market very quickly will take this invention into account and will modify its behavior making this  miracle system useless.

This is the extreme variation of this theory, that does not leave a human any chance  to make money on the stock market consistently.


It is not too bad ...

This is a basic rule that we should accept: the rules are constantly changing - the stock market is efficient. However, in reality the situation is not so bad.

Suppose somewhere in the World some guy really invents some miracle system. And he really makes money with it, big money. The stock market needs some time to reveal his system. Meanwhile, the stock market is huge, and many people lose their money there. So, his majesty Stock Market can allow our guy to make money: the smartest wins, the stock market simply redistributes the wealth. The only thing that Stock Market cannot allow to this guy is getting his profit forever. What conclusion can we make from this? Our smart guy should invent something else. You see, the stock market does not provide a whole life warranty. It simply has a lag that is enough for some person to become rich using some method.

What does this idea give to us? Practically speaking, for traders and investors in 1920 - 1930s it was possible to make money on the stock market using the simplest moving average crossover (this is the technique available now practically in any program). But in the end of 1940s when moving averages started being used widely (military guys brought this idea at the end of WW2) the stock market took it into account. The advantage for some became a common thing. Now it is used for educational purposes mostly (see R. Colby's "The encyclopedia of technical market indicators"; in this book book it is shown that within 1915-1937 years it was possible to make 52% of annual profit using moving averages strategy, while in 1937-1992 years this profit has dropped to 10%).

This is another example: somehow, from the middle of the year 2007 till now (the beginning of 2010), the 117 calendar days cycle has been very strong. It has not been so before 2007, and it is weakening now. This time span exactly covers the most recent Great Financial Crisis. You can see it from this wavelet diagram; a horizontal red stripe there indicates that this cycle has been very active:

This is how this cycle has worked in time: 

It was possible to reveal this cycle in the end of year 2008 and use it as a trading tool. It became obvious then, at least, for some. In the end of 2009, it became obvious for many and lost its power. (BTW, this cycle exactly corresponds to Mercury synodic cycle.)

Our goal is to find the "inefficiency window" for some idea, the period when some working patterns/models are not covered by the stock market yet. If some theory becomes widely known, this a sign that this theory does not work anymore. Stock market always gives us the chance, though it does not allow us to sleep.   


So,  I do not provide the software that allows to reveal some hidden pattern in the stock market that will work forever. I do not do that. I provide the software that allows to deal with constantly changing market.

The most important things you should do:

- learn to think and analyze;

- learn always to study something new;

- learn to leave behind the model that has worked well before, but does not work now;

- learn to give not too much time to post-factum explanations.

We provide the ideal environment for that. This is the only philosopher stone I know.


February 28, 2010

Toronto, Canada

PS. written while Canada was playing with the USA for the Olympic Gold in hockey. Canada won!