Все знают народную примету о том что американский stock market надо покупать в октябре и продавать (close positions) in May. Let's check this sign. Let's take the DJ-30 index and test it over a long period of time, for example, behind 100 of recent years. We look at the distribution of profitability by months.
For the last 100 years:
We see that, really, after May the problem period begins. July only exception — profitable. Of course, there are doubts that such an alignment can be obtained due to remote years, when the market was different. Therefore, you can discard older years and test on a segment in the last 50 years.
For the last 50 years:
We see that the situation, exactly identical. But for clarity, we discard half of the years and test on more recent data in 25 of recent years.
For the last 25 years:
We see what for 25 years nothing has changed. Well, maybe on the most recent data for the last 10 years the situation will be different?
For the last 10 years:
Not, practically, the trend remains the same with small, unprincipled changes.
It turns out that really, sign `` sell in may" occurs?
Let's make one more check — let's imagine that we are in 1990 year and do the same test with 1917 on 1990 year.
With 1917 year on 1990 year:
And what do we see — June and August are already profitable months, and October is unprofitable!!! I.e, the whole pattern collapses. It turns out, if we were in 1990 year, then they would not know that in the next decades it is necessary to sell in May and buy in October, since no one knows the future except psychics and clairvoyants.
And let's also imagine that we are in 2000 year and we have DJ-30 data only for the last 10 years.
With 1990 year on 2000 year:
We see that there was no folk omen here either, `` sell in May" and no one could guess about her for sure, since there could only be one sign — & quot; sell in august" :)
