Bookshelves of Russian shops filled with books of both Russian and foreign authors on the successful sale of the stock markets, commodities markets, and FOREX. The authors provide a great number of various boards (free if you do not take into account the cost of books), how to learn to understand the market characteristics of its motion, and quickly become a successful trader. In the course of all is: the science and mysticism.
In 2000, on the shelves of Russian stores have the book, Bill Williams, entitled "New Dimensions in the exchange trade.
Here is an excerpt from the book: "My (Mr. Williams), the trade group, after fifteen years research has begun to use the science of chaos, along with quantum physics, holography, cybernetics, nonlinear dynamics, information theory and fractal geometry in the world stock and commodity transactions" . The impressive start, is not it?
Trading System Investor's Dream
Fifteen studies Mr Williams and Co. formed the basis for the trading strategy, which is called the "investor's dream" (Investor's Dream).
However, according to developers, allowing to implement this vision? The author of the book calls them "five zagovorennyh bullets."
These are:
1. Fractal
2. Amazing oscillator
3. Market acceleration
4. Line balance
5. Zone
In addition, there is a filter signals and four way closure of transactions:
1. The red line
2. Five the other bars the same color
3. Green Line
4. The signal in the opposite direction.
You can select any of five methods for generating signals or use them all together for an aggressive game - serial additions of transactions, the signals which are generated by each of the 5 'zagovorennyh bullets. "
According to the author of the book, its trading system works very well in any market, allowing the trader to choose the most opportune moment to enter the market and exit from it. The book is replete with examples of how well each works from five zagovorennyh bullets, "but, unfortunately, mostly in the ascending price ranges.
In the book, little attention was paid to the choice of these trend plots. According to the author's strategy, the filter built into the system, will help to enter the market and leave it at the most opportune moment.
The filter is a three simple moving averages:
13-ty periodnoe moving average, shifted forward by 8 bars;
8 and periodnoe moving average shifted forward by 5 bars;
5-periodnoe moving average shifted to the 3 bars ahead.
Mutual location of these three lines to one another, as well as fractal signal on one of these lines and determine the opportune moment to start the game.
In addition, the author of the book surpassed my attention, an important question: how and how much stock you need to buy or sell. That is, the program does not analyze what the most probable amount of the bought / sold the lot on the traded asset. But this issue is directly related to liquidity risk. Moreover, it is especially important with the "aggressive" strategy game - a sequential adding more and more open positions.
You may ask, how is this a dream? Yes just did - $ 3,400.
Unfortunately, the demo version of Investor's Dream can not assess the effectiveness of the strategy. This article will not detail all the steps of the strategy (this is the book). Let's just one of the 5 so-called "zagovorennyh bullets" - Trade through fractal.
Fractals and trading fractals
Generally, the concept of fractal is not appropriate to use in the book "New Dimensions in the exchange trade." Strict definition of fractal, since the opening of the fractal structures of complex natural objects Mandelbrot, no. Furthermore, this fractal structure, repeated at different time scales. If we consider the price series, at the minute, hour, and so on price charts. The book is, on different time scales is silent.
It is therefore not in this article use the word fractal, and replacing it with the word pattern (pattern, pattern, structure), which is most appropriate.
So, the pattern is a series of at least 5 consecutive price bars in which the highest peak in front and behind him are two bars with lower maxima. Let's call this pattern the structure of the first type - pattern1. Similarly, the pattern of the second type - pattern2 is a series of at least 5 consecutive price bars in which to the lowest minimum and it is for two bars with higher minimums.
Figure 1 presents the types of patterns.
Models A and D represent the patterns of the first type - pattern1, while model A consists of 5 bars, a model D of 6. Models B and C are both patterns of both types. Bars, which are drawn the cap (in the presentation of the color of top hat - a green hat down - red) are called base price bar patterns.
Some authors offer a more sophisticated model of the formation of patterns, such as presented in Figure 2 and Figure 3. Letter F in the figures marked the basic price bars.
Models of the first type of pattern - pattern1
Why Patterns?
Why such a structure, as a pattern, was chosen as a tool for the game on the market?
An analysis of price charts you can see that if the graph, a pattern of first type (see Figure 4), the base price bar is:
1. a point of correction (rollback), after which the quotations will continue to grow,
2. a point and a change of trend.
Further, in the figures above the base price bar patterns of the first type is a green triangle, and for patterns of the second type of basic price bar is the red triangle.
Figure 4 arrows with inscriptions Correction and Turn identifies patterns of the first type, after which the correction occurred and spread, respectively.
Similar to the pattern of the second type (see Figure 5):
1. a point of correction (rollback), after which the quotations will continue to fall,
2. a point and a change of trend.
In Figure 5 arrows with inscriptions Correction and Turn identifies patterns of the second type, which occurred after the correction and turn.
Signal patterns
Once the pattern of any type of shape, its base price bar signal (order) to buy or sell.
The signal pattern for purchase
Once a pattern emerged the first type, exhibited stop-buy order at a minimum change in price (currently at a minimal change in stock market prices is $ 0.01) higher than its base price High price bar.
The signal pattern for sale
Once formed the pattern of the second type exhibited stop-sell order at a minimum price change is lower than its base price Low price bar.
Regulation of signal patterns
1. Alerts patterns on the purchase / sale generated independently from each other and are valid until either does not work, or until the pattern of the same type do not generate a new, fresh tone. In doing so, the old signal, if it did not work, replaced by a new signal. This provides more favorable conditions (in terms of prices) for the purchase / sale.
2. Order to buy / sell at the time of operation must be above / below the line 8? Mi periodnogo rolling average, based on averages of price bars, and shifted to the 5-y periods ahead.
3. One of the options for closure of all previously discovered long / short positions: close all previously open long / short positions occur on the closing price of the bar when its closing price is below / above the line 8? Mi periodnogo the rolling average, based on averages of price bars, and shifted to the 5-y periods ahead.
Emulation and the results
While developers Strategy Investor's Dream, and pay attention to the fact that it can not be regarded as a mechanical trading system in the full sense of the word, however, we have a simulation game on the patterns, in order to determine whether the rule is well on the patterns of trade, According to the author the book.
For this stock the most liquid companies, the average volume of trades for which the period from 01/05/1998 to 11/19/2002 was not less than 500,000 shares (liquidity) - total 296 companies. In doing so, we are not interested in price movements of selected companies, as took for granted the view of the author's strategy of Investor's Dream, that the filter (three moving averages) will help determine the best point of entry and exit into the market regardless of price dynamics.
Game emulate each company separately on long positions, short positions and the possibility of opening of both types of positions. This was done to determine how balanced on the type and position of the company (or rather price dynamics) strategy game patterns.
Emulation parameters were as follows:
1. The sum accounts $ 25000 $ 50000 and $ 75000. These amounts are typical for the private player on the Russian stock market.
2. The amount of the account is constant (in the case of a loss, you need to update your account until the original amount, in case of win - win was withdrawn from the account).
3. Number of shares, each time buying / selling in one transaction (ranged from 100 to 2000 in steps of 100).
4. An aggressive strategy of the game - you can build on the position - buy with open long positions and sell in the open short position.
5. Emulation was conducted assuming the absence of margin accounts.
6. Transaction costs - feature volume purchased / sold the lot.
Results emulation:
The game in a long
Table 1 shows the distribution of probability density functions of total return on equity as a result of emulation of the game only in length.
From the figures shows that:
All distribution, regardless of the size of the lot and the amount of the game have heavy right tails that are characterized by "large" values of yield. With the increase of volume of lots, and the amount of the game, the right tail thicken.
Maximum density distribution function is in the negative area for the small lots, and any amount in the game. When increasing amounts of lots, the maximum shifts in the positive region, while remaining fairly low value.
When the amount of $ 25000 in the game and the volume of lots 1500 and 2000 shares of the density function has the form of delta-functions (with a maximum at zero) with a thick right tail, which may mean that the profitability of most of the 296 companies in emuliruemy period was approximately zero.
The game in brief
Table 2 presents the distribution of probability density functions of total return on equity as a result of emulation of the game in short.
The distributions have thick tails right and left, depending on the amount of game and size of lots. With the increase of volume of lots, and the amount of the game, the right tail thicken.
Maximum density distribution function is in the negative domain for any volume of lots and any amount in the game.
When the amount of $ 25000 in the game and the volume of lots 1500 and 2000 shares of the density function have the form of Delta-function (with a maximum at zero) with a thick left and right tails, which may mean that the profitability of most of the 296 companies in emuliruemy the period was approximately zero.
Playing in the long and short
Table 3 presents the density function probability distribution of total return on equity as a result of emulation of the game, you can open as long or short position.
1. The distributions have thick tails right and left, depending on the amount of game and size of lots. With the increase of volume of lots, and the amount of the game, the right tail thicken.
2. Maximum density distribution function is in the negative area for the small lots, and any amount in the game. When increasing amounts of lots, the maximum is shifted to the region of zero, while remaining fairly low value.
Conclusions
Ironically, the aggressive strategy of the game, which in our case is only possible for small quantities of lots, does not lead to a substantial increase in profitability. Most of the shares in the small quantities of lots and any amount left in the game in a negative, and only a small fraction reached the "plus".
But when playing for long amounts Chiva to increase the number of open positions, increasing the likelihood that some of them will be closed with the profit, which ultimately lead us to the plus (this aggressive strategy is called piramiditsey and, strictly speaking, leads to profits with unlimited amount of money in the account and a large interval of time).
With the increase of the same volume with a limited amount of lots in the game, when the possibility of opening a consistent number of transactions running decreases (because of restrictions the total bill), the total value or returns are large enough, or the game leads to significant losses.
Big gains were shown on the companies, price dynamics are characterized by prolonged periods of rising or falling (see Figure 6) - is himself and a filter based on a combination of three moving averages and patterns.
However, comparable results could be obtained, and when playing on two moving averages. As is known, a strategy game on moving averages show good results on distinct trend of price dynamics.
For companies, the price dynamics are replete with "frequent" periods of recession and recovery, and volatility was high, the game resulted in a loss (see Figure 7). Ie filter designed to identify optimal points of entry and exit into the market failed in their task: either when coming out of the market (and the moving averages always signal with delay of completion of the trend) were so badly that the entire profit of "eat", or the parameters of moving averages were not optimal.
Of all of the above, we can conclude that, unfortunately, Investor's Dream can not be described as a dream partner.
The strategy has proved not balanced, as the type of game (for long positions more positive returns for short - a large negative returns) and the ensemble companies.
The problem of the choice of assets and the best moments of the entrance / exit into the market, which in the opinion of the author of the book and the strategy Investor's Dream, should be resolved through the filter of the three moving averages are not resolved. In addition, the authors state that they found in the hard work of filtering (see above), suitable for any market and any price dynamics, which turned out not true.
All the above described observations once again confirm the infallible truth: the market is not a hospital and requires a constant attention.
Optimal parameters for the history of the strategy does not guarantee a profitable game in the future. It may be that the market situation will require the conversion not only the parameters of certain indicators, but changing the optimization criterion.
In 2000, on the shelves of Russian stores have the book, Bill Williams, entitled "New Dimensions in the exchange trade.
Here is an excerpt from the book: "My (Mr. Williams), the trade group, after fifteen years research has begun to use the science of chaos, along with quantum physics, holography, cybernetics, nonlinear dynamics, information theory and fractal geometry in the world stock and commodity transactions" . The impressive start, is not it?
Trading System Investor's Dream
Fifteen studies Mr Williams and Co. formed the basis for the trading strategy, which is called the "investor's dream" (Investor's Dream).
However, according to developers, allowing to implement this vision? The author of the book calls them "five zagovorennyh bullets."
These are:
1. Fractal
2. Amazing oscillator
3. Market acceleration
4. Line balance
5. Zone
In addition, there is a filter signals and four way closure of transactions:
1. The red line
2. Five the other bars the same color
3. Green Line
4. The signal in the opposite direction.
You can select any of five methods for generating signals or use them all together for an aggressive game - serial additions of transactions, the signals which are generated by each of the 5 'zagovorennyh bullets. "
According to the author of the book, its trading system works very well in any market, allowing the trader to choose the most opportune moment to enter the market and exit from it. The book is replete with examples of how well each works from five zagovorennyh bullets, "but, unfortunately, mostly in the ascending price ranges.
In the book, little attention was paid to the choice of these trend plots. According to the author's strategy, the filter built into the system, will help to enter the market and leave it at the most opportune moment.
The filter is a three simple moving averages:
13-ty periodnoe moving average, shifted forward by 8 bars;
8 and periodnoe moving average shifted forward by 5 bars;
5-periodnoe moving average shifted to the 3 bars ahead.
Mutual location of these three lines to one another, as well as fractal signal on one of these lines and determine the opportune moment to start the game.
In addition, the author of the book surpassed my attention, an important question: how and how much stock you need to buy or sell. That is, the program does not analyze what the most probable amount of the bought / sold the lot on the traded asset. But this issue is directly related to liquidity risk. Moreover, it is especially important with the "aggressive" strategy game - a sequential adding more and more open positions.
You may ask, how is this a dream? Yes just did - $ 3,400.
Unfortunately, the demo version of Investor's Dream can not assess the effectiveness of the strategy. This article will not detail all the steps of the strategy (this is the book). Let's just one of the 5 so-called "zagovorennyh bullets" - Trade through fractal.
Fractals and trading fractals
Generally, the concept of fractal is not appropriate to use in the book "New Dimensions in the exchange trade." Strict definition of fractal, since the opening of the fractal structures of complex natural objects Mandelbrot, no. Furthermore, this fractal structure, repeated at different time scales. If we consider the price series, at the minute, hour, and so on price charts. The book is, on different time scales is silent.
It is therefore not in this article use the word fractal, and replacing it with the word pattern (pattern, pattern, structure), which is most appropriate.
So, the pattern is a series of at least 5 consecutive price bars in which the highest peak in front and behind him are two bars with lower maxima. Let's call this pattern the structure of the first type - pattern1. Similarly, the pattern of the second type - pattern2 is a series of at least 5 consecutive price bars in which to the lowest minimum and it is for two bars with higher minimums.
Figure 1 presents the types of patterns.
Models A and D represent the patterns of the first type - pattern1, while model A consists of 5 bars, a model D of 6. Models B and C are both patterns of both types. Bars, which are drawn the cap (in the presentation of the color of top hat - a green hat down - red) are called base price bar patterns.Some authors offer a more sophisticated model of the formation of patterns, such as presented in Figure 2 and Figure 3. Letter F in the figures marked the basic price bars.
Models of the first type of pattern - pattern1
Why Patterns?
Why such a structure, as a pattern, was chosen as a tool for the game on the market?
An analysis of price charts you can see that if the graph, a pattern of first type (see Figure 4), the base price bar is:
1. a point of correction (rollback), after which the quotations will continue to grow,
2. a point and a change of trend.
Further, in the figures above the base price bar patterns of the first type is a green triangle, and for patterns of the second type of basic price bar is the red triangle.
Figure 4 arrows with inscriptions Correction and Turn identifies patterns of the first type, after which the correction occurred and spread, respectively.Similar to the pattern of the second type (see Figure 5):
1. a point of correction (rollback), after which the quotations will continue to fall,
2. a point and a change of trend.
In Figure 5 arrows with inscriptions Correction and Turn identifies patterns of the second type, which occurred after the correction and turn.Signal patterns
Once the pattern of any type of shape, its base price bar signal (order) to buy or sell.
The signal pattern for purchase
Once a pattern emerged the first type, exhibited stop-buy order at a minimum change in price (currently at a minimal change in stock market prices is $ 0.01) higher than its base price High price bar.
The signal pattern for sale
Once formed the pattern of the second type exhibited stop-sell order at a minimum price change is lower than its base price Low price bar.
Regulation of signal patterns
1. Alerts patterns on the purchase / sale generated independently from each other and are valid until either does not work, or until the pattern of the same type do not generate a new, fresh tone. In doing so, the old signal, if it did not work, replaced by a new signal. This provides more favorable conditions (in terms of prices) for the purchase / sale.
2. Order to buy / sell at the time of operation must be above / below the line 8? Mi periodnogo rolling average, based on averages of price bars, and shifted to the 5-y periods ahead.
3. One of the options for closure of all previously discovered long / short positions: close all previously open long / short positions occur on the closing price of the bar when its closing price is below / above the line 8? Mi periodnogo the rolling average, based on averages of price bars, and shifted to the 5-y periods ahead.
Emulation and the results
While developers Strategy Investor's Dream, and pay attention to the fact that it can not be regarded as a mechanical trading system in the full sense of the word, however, we have a simulation game on the patterns, in order to determine whether the rule is well on the patterns of trade, According to the author the book.
For this stock the most liquid companies, the average volume of trades for which the period from 01/05/1998 to 11/19/2002 was not less than 500,000 shares (liquidity) - total 296 companies. In doing so, we are not interested in price movements of selected companies, as took for granted the view of the author's strategy of Investor's Dream, that the filter (three moving averages) will help determine the best point of entry and exit into the market regardless of price dynamics.
Game emulate each company separately on long positions, short positions and the possibility of opening of both types of positions. This was done to determine how balanced on the type and position of the company (or rather price dynamics) strategy game patterns.
Emulation parameters were as follows:
1. The sum accounts $ 25000 $ 50000 and $ 75000. These amounts are typical for the private player on the Russian stock market.
2. The amount of the account is constant (in the case of a loss, you need to update your account until the original amount, in case of win - win was withdrawn from the account).
3. Number of shares, each time buying / selling in one transaction (ranged from 100 to 2000 in steps of 100).
4. An aggressive strategy of the game - you can build on the position - buy with open long positions and sell in the open short position.
5. Emulation was conducted assuming the absence of margin accounts.
6. Transaction costs - feature volume purchased / sold the lot.
Results emulation:
The game in a long
Table 1 shows the distribution of probability density functions of total return on equity as a result of emulation of the game only in length.
From the figures shows that:
All distribution, regardless of the size of the lot and the amount of the game have heavy right tails that are characterized by "large" values of yield. With the increase of volume of lots, and the amount of the game, the right tail thicken.
Maximum density distribution function is in the negative area for the small lots, and any amount in the game. When increasing amounts of lots, the maximum shifts in the positive region, while remaining fairly low value.
When the amount of $ 25000 in the game and the volume of lots 1500 and 2000 shares of the density function has the form of delta-functions (with a maximum at zero) with a thick right tail, which may mean that the profitability of most of the 296 companies in emuliruemy period was approximately zero.
The game in brief
Table 2 presents the distribution of probability density functions of total return on equity as a result of emulation of the game in short.
The distributions have thick tails right and left, depending on the amount of game and size of lots. With the increase of volume of lots, and the amount of the game, the right tail thicken.Maximum density distribution function is in the negative domain for any volume of lots and any amount in the game.
When the amount of $ 25000 in the game and the volume of lots 1500 and 2000 shares of the density function have the form of Delta-function (with a maximum at zero) with a thick left and right tails, which may mean that the profitability of most of the 296 companies in emuliruemy the period was approximately zero.
Playing in the long and short
Table 3 presents the density function probability distribution of total return on equity as a result of emulation of the game, you can open as long or short position.
1. The distributions have thick tails right and left, depending on the amount of game and size of lots. With the increase of volume of lots, and the amount of the game, the right tail thicken.2. Maximum density distribution function is in the negative area for the small lots, and any amount in the game. When increasing amounts of lots, the maximum is shifted to the region of zero, while remaining fairly low value.
Conclusions
Ironically, the aggressive strategy of the game, which in our case is only possible for small quantities of lots, does not lead to a substantial increase in profitability. Most of the shares in the small quantities of lots and any amount left in the game in a negative, and only a small fraction reached the "plus".
But when playing for long amounts Chiva to increase the number of open positions, increasing the likelihood that some of them will be closed with the profit, which ultimately lead us to the plus (this aggressive strategy is called piramiditsey and, strictly speaking, leads to profits with unlimited amount of money in the account and a large interval of time).
With the increase of the same volume with a limited amount of lots in the game, when the possibility of opening a consistent number of transactions running decreases (because of restrictions the total bill), the total value or returns are large enough, or the game leads to significant losses.
Big gains were shown on the companies, price dynamics are characterized by prolonged periods of rising or falling (see Figure 6) - is himself and a filter based on a combination of three moving averages and patterns.
However, comparable results could be obtained, and when playing on two moving averages. As is known, a strategy game on moving averages show good results on distinct trend of price dynamics.
For companies, the price dynamics are replete with "frequent" periods of recession and recovery, and volatility was high, the game resulted in a loss (see Figure 7). Ie filter designed to identify optimal points of entry and exit into the market failed in their task: either when coming out of the market (and the moving averages always signal with delay of completion of the trend) were so badly that the entire profit of "eat", or the parameters of moving averages were not optimal.
Of all of the above, we can conclude that, unfortunately, Investor's Dream can not be described as a dream partner.The strategy has proved not balanced, as the type of game (for long positions more positive returns for short - a large negative returns) and the ensemble companies.
The problem of the choice of assets and the best moments of the entrance / exit into the market, which in the opinion of the author of the book and the strategy Investor's Dream, should be resolved through the filter of the three moving averages are not resolved. In addition, the authors state that they found in the hard work of filtering (see above), suitable for any market and any price dynamics, which turned out not true.
All the above described observations once again confirm the infallible truth: the market is not a hospital and requires a constant attention.
Optimal parameters for the history of the strategy does not guarantee a profitable game in the future. It may be that the market situation will require the conversion not only the parameters of certain indicators, but changing the optimization criterion.
Franklin & Grant
No comments:
Post a Comment