Not knowing peace and rest, in moonlight and sunlight, we make money out of thin air in order to waste it again. (With). Knowing neither sleep nor rest, in moonlight and sunlight, we make money out of thin air in order to blow it down the drain again...

I do not know to whom this statement belongs, but it definitely fits the topic of our conversation today. But first, I'll tell you a little about myself. I am an investor with a little over a year of experience. I invest not only in, but also in various investment companies, funds, real. I can't get by without hype. In general, everything is going steadily, capital is growing. Previously, I have already participated in similar competitions more than once, and watched this one for several weeks. In the end, I decided to share with you one of the most unusual ways to make money. Its main essence is that you can earn both on investing your funds and without spending a single penny.

In total, I plan to write 4 articles, including this one, and fully unlock the potential of the investment instrument I propose. This tool is known to most of you, moreover, many of you use it daily, but for other purposes. We will talk about a group (community, public) VKontakte. The question arises - since when has the VKontakte group become an investment tool? The answer is simple: you are investing your time and/or money to make a profit.

I will immediately make a few remarks. Firstly, this series of articles will be narrowly focused, but by analogy, you can easily use the information received to create a group in any other social networks. Perhaps some of the information can be used to work on your site. Secondly, I am not an SEO specialist and I have not studied programming either, however, I know my subject quite well. Thirdly, I have no desire to write frank instructions for dummies, so we will miss simple moments, focusing on more important things. And fourth, below I will present my own experience, with my own examples, so a big request - if you want to criticize, show what you have achieved.

Today, according to the plan, we have an introductory part, so to speak, just to show that the current TOPs of the competition will soon have another rival. Well, to tell briefly what we will consider in the future. So, let's go over the plan.

Plan:

1. Create a group. In this article, we will consider the main points that need to be considered when creating a VKontakte group. Let me give you a few personal examples.
2. Group promotion. Here we will talk about ways to promote the group. Again, there are plenty of examples.
3. Group income. In conclusion, we will talk about how to monetize the efforts expended and start receiving a stable income. Again, I will give a few examples.

In custody:

In conclusion, I want to note once again that the entire series of articles will be based on

Not knowing peace and rest,
In moonlight and sunlight
I make money out of thin air
To immediately let them go to the wind
© I. Guberman (at least he thinks so)

While the robot was working hard during the day, I managed to run 5 km, sleep twice, remove redundant settings of the robot parameters and started combining filters of different types and orders in one system. And given the fact that I slept well during the day, I worked quietly all night. And only at 7 in the morning I went to sleep for 3 hours, which was quite enough ... Now I will finish this text, and again for a walk-jog.

I'm afraid that my regular readers will soon start to get a little crazy. They (readers) are already accustomed to the fact that stochastic trends give some kind of constant picture and have completely lost sight of my remark that there can be an infinite number of combinations of such trends. And in August, "Ostap suffered" and the number of simplifications, mergers and various modifications can no longer be counted. Of the pluses, there is only one thing - fewer parameters that need to be configured, an increasing degree of autonomy in the actions of the trading robot.

Now once again about stochastic wave trends. Using the principles inherent in the SWT method, you can build an infinite set of bandpass filter sets of different orders, different steps, designed by different methods and having different positions of poles and zeros on the complex plane of the z-transform (I scare this a little with scary terms, behind which there is no particular difficulty, however, there are many alternatives. There can be many filter sets, and the trends for these sets will be different, but the principles of working with waves remain the same. One of my acquaintances, whose method I deciphered from a couple of words in correspondence, uses a similar approach, which, translated into the language of the SWT method, is equivalent to a filter comb step of 1.618 (Fibonacci number). Implementing a similar principle within the framework of the SWT method, I drowned in waves, there were too many of them and there was no special point in this, since they were very strongly correlated. 9 wave trends of the SWT method are also quite a few, but it is still 4-5 times less than the filter comb with a step of 1.618 gives.

But back to filters.
At one time, I experimented with different designs, of which only three have survived to this day for various reasons:
- a standard simplest set of second-order filters obtained by converting an analog prototype into digital form by switching from differential equations to equations in finite differences, which is characteristic of the time series analysis apparatus;
- a filter obtained by the bilinear z-transform method, which differs from the previous one. that there is no aliasing effect due to the fact that periodic processes with a period that is a multiple of the timeframe interval in a discrete form become indistinguishable;
- well, and a fourth-order filter, although there was no particular need for it, but just to see what happens when the filtering mode is tightened.

Now I have combined all three filters into one software box with filtering mode switching. This is done both in the indicators and in the robot, which allows, during tests in one pass, to scan not only the type of adaptive mode and vectors of typical settings for trend configuration, but also to select the type of filter and the degree of division of the price chart into a set of trends.

The visual differences between the original prototype filter built according to the finite difference equation and the filter implemented using the bilinear z-transform method are small.
In the latter case, the waves move a little more smoothly. which can be seen in the presented graph. Prototype filter waves below. Translating the results into the usual language of the trader - a little less "false" positives. Although there is nothing false in the market, except for our expectations. And the market is always right.

There is a larger difference between the results of fourth-order (upper indicator) and second-order (lower indicator) filters. The course of trends is even smoother, the inertia is higher, faster movements are shifted towards shorter trends, reducing the noise characteristics of long-term trends.

I am not going to dig deep into the analysis using fourth-order filters, but in the robot, if such a need arises and there are corresponding results, I will also use this setting.

In the meantime, for comparison, the test results for three options.

1. Second order filter - initial prototype.

2. Second order filter - bilinear z-transform.
More profit, slightly less drawdown. The stats are better.

3. Fourth order filter.
The profit is smaller, there are fewer transactions, but the equity curve is smoother and the statistics allows you to compensate for the lack of profit by increasing the volume of transactions.
As you can see, during the testing interval, the robot did not trade sells at all in this mode and was only in long positions.
The previous versions also did not indulge in sales, since the trends of the senior level of the hierarchy are in the growth phase, but there were still rare sales on corrections.

Here are the preliminary results.

Let's see what to do with it. Life will show.

P.S. Yes, another test in pips to eliminate the heresy of reinvestment and get the result of a pure trading algorithm.

For a little over 4 months, the robot has threshed 68388 pips of profit.
Average MO trade +83 pips.
Average profitable trade 315 pips.
Average losing trade -219 pips.
The percentage of profitable trades is 56.59%.

When transferring money down the drain, please take into account one indisputable truth - a person is able to learn from his mistakes. Only on their own! And that's all ... provided that you are ready to learn. In other cases, the market simply takes your money and gives it to someone for more exemplary behavior.

And so it happened on Monday, when Seryozha tried to catch a “rebound”. And to be honest, it was a game of pure luck. Almost like a casino :)

With the conclusions, the situation looks quite simple - NEVER try to catch such things again. Well, fuck them :) And second, no less important - being a bull in a bear market is a very uncomfortable state. The feeling inside is like boots are tight, only all over the body. Well approximately...

I liked this picture better.

He returned the depot to its previous form by adding 166 rubles to the profit. Boo-ha-ha :)) And that - 9.666. The beginning, let me remind you - 9.500.

The most important thing (for me personally) is the positive state of the overall balance, and not of each specific transaction. Although, by the fact of being in the market, from time to time you catch yourself thinking that you always want profits. Damn - ALWAYS! It's like a goddamn drug that can hit you like the ultimate dope.

At the moment, 1.996 + 1.735 = 3.731 rubles have been spent on training (in total). The conclusions were made immeasurably, including about our own unreasonable behavior, but this information will manifest itself a little later, but for now we will continue to survive :)