Design trading system using method of random input

Design a trading system using the method of random input.

Design a trading system, a standard way.

Usually the design of the trading system begins with the entrance of the transaction signal. Long-term observation of the price chart gives the trader to identify certain patterns, which allows to profit from the market. For example, this may be a signal from the indicator.

the input signal is modeled as a program and added to the output signal of the transaction, e.g., at the end of the working day. Start creating a trading system is necessary.

Typically, the shape of the yield curve is not satisfied with the developer. For the best shape of the yield curve added to the system:

  • the trend filter
  • trend strength filter
  • confirmation of the input signal
  • output signal
  • filter output signal
  • Stop
  • and so on

At each stage, there are more options for optimization. The more, the easier it is to select a combination of options that bring the shape of the yield curve, based on historical data, the ideal.

As a result, after carrying out a plurality of tests, the yield curve becomes approximately the following form:

Design trading system using method of random input

The system starts to work and brings disappointment

Design trading system using method of random input

“Body kit” system with additional capabilities to control the trade entry and exit, in any case, would mean fit to the historical data, and limit the lifetime of the trading system.

Design a trading system using the method of random input.

article
“Random input – an effective method of testing trading systems”
detail the testing equipment units trading system using this method.

The idea is to use a “kit” of the trading system to enhance its real effectiveness and not to fit to the historical data.

In the case of this approach, the design of the trading system does not begin with the input signal, and to determine the required functionality.

For example, it is imperative for the existence of the trend system:

  • Filter to the input of a transaction in the direction of the main trend.
  • Trailing stop.
  • Signal at the output, in the case of a reversal of the trend (there may be several).

Each of these units can be tested separately using random input method. This uses standard parameters used for indicators. Details are in the following articles:

  • “Test Pattern” external bar “at random entry".
  • “Choosing the best trailing stop using the method of random input and position tracking system MultiStop_Pro".
  • “Selection of the optimum signal output from the transaction using the method of random input and system maintenance position MultiStop_Pro".
  • “Selection of the optimal display of the trend by the method of random input".

The result is a “kit”, not tied to a unique input signal, and to traded instruments and timeframe. Given that all service units of the trading system have a positive expectation, the success of creating a stable trading system increases.

Only after testing the system components, you can go to find the optimal signal at the entrance to the deal.

Selection of the optimal signal to the input of a transaction.

In the articles mentioned above, testing was conducted trading system blocks for futures pair Ruble dollar, Si-6.16 contract. The following results were obtained from the results of testing by random inputs:

  • The optimum filter at the entrance to the trade in the direction of the main trend – two exponential moving averages timeframe Day filter periods standard 9 and 18.
  • Trailing Stop – a classic trailing stop on the current timeframe, parameters calculated automatically on the basis of collateral.
  • output signals:
    • Output on an external bar, timeframe Day;
    • Entering the trend reversal, determined by the divergence of price and MACD on timeframe H1.

It remains to choose the optimal signal to the input of a transaction. trading platform «BetaRelatedMarkets_V3» used for testing.
We examine the following signals to open the position:

  • The indicator goes out of overbought oversold zone.
  • Indicator signal crosses below or above pereprodannosti prekuplennosti level.
  • Indicator crosses the median. For indicators without clearly defined levels median = 0, with indicators for the levels, median = (level + level perekuplennosti pereprodannosti) / 2
  • Turn indicator to 3 points below or above pereprodannosti perekuplennosti level. Determined by analyzing the 3 neighboring bars.
  • Turn indicator on 5 points below or above pereprodannosti perekuplennosti level. Determined by analyzing the 5 neighboring bars.
  • Key reversal in the chart below pereprodannosti prekuplennosti level or higher.
  • Divergence or divergence indicator values ​​and prices. Indicator values ​​should be below or above pereprodannosti perekuplennosti level.
  • The indicator is adjusted by max / min. Indicator values ​​should be below or above pereprodannosti perekuplennosti level. The indicator should adjust the achieved extremum on the threshold.

46 oscillators to determine the point of entry into the transaction available to the trading platform.

Schedule optimization is as follows:

Design trading system using method of random input

As can be seen from the chart, 95% of the results are in the positive area. That is, the signal at the entrance to the deal does not matter all that much, as it seems at first glance.

We select the input signal for a transaction indicator divergence Rates% b.

Bektest system:

Design trading system using method of random input

Schedule system yields:

Design trading system using method of random input

It remains to check the system for other contracts, such as Si-3.16.

Bektest system:

Design trading system using method of random input

Schedule system yields:

Design trading system using method of random input

conclusions

The use of random input allows you to design trading systems with a minimum of parameters.
The probability of “fit” under the historical data is much lower, as the blocks are tested separately on random sequences.
All who are interested in this method of designing trading systems invite a discussion on this topic.

Related posts

Next posts

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

w

Connecting to %s