Design method of trading model

The main reference here is the classification method of all kinds of relevant information, will be divided into the following three types of models: technical analysis of the trading model, basic analysis of the trading model, mathematical measurement of the trading model.

1, technical analysis trading model

Technical analysis trading model refers to the use of market trading data such as opening price, closing price, turnover, and so on, and through the computer trading indicators, after a systematic search and testing and optimization of the trading model, the theoretical basis of which is mainly built on the basis of existing traditional technical investment theories, such as graphical analysis, averaging theory, etc., and through a large number of statistical analysis and testing. A large number of statistical analysis and testing. The biggest advantage of this model is that it eliminates the influence of investor's emotions in trading decisions, especially the subjectivity and blindness in judging major events; avoids analytical errors due to information asymmetry; ensures consistency in trading analysis; and provides investors with a method of risk control.

The following discussion focuses on three trading models in the technical analysis trading model:

1) Trading model based on graphical pattern recognition

This type of model is mainly based on the traditional classic graphs such as the head and shoulders tops, double bottoms, triangles, etc., to market trend capture, to build a position in the trading system. But in practice, it also has many problems: risk control, like the head and shoulders top, double bottom, triangles and other trading charts, according to the traditional trading views, investment risk / reward ratio is generally 1:1, the actual battle managers will face a huge risk of the net value of the fund; analysis of the majority of subjective judgment is based on the lack of objective judgment standards; the current increase in the domestic futures market, technical analysis of users, resulting in the classic charts Forms of false signals and the subsequent increase; classic foreign chart analysis theory in the country there is a considerable difference; lack of statistical data.

2) Trend-tracking based trading models

This type of model is mainly based on the designer's data statistics, capturing the turning point of the price, and then assuming that the trend will continue, and according to the direction of the trend to build a position in the trading system, such as MACD, SAR, moving averages and so on. The trading model is characterized by not buying at the lowest price and not selling at the highest price, giving up the profit of the section before and after the market, and the profit mainly comes from capturing the middle part of a big wave of the market. Its ability to capture the turning point of the market according to the designers of the design of the sensitivity of different, strong sensitivity of the trading model on the trend reversal reaction quickly, but also more false signals; low sensitivity of the trading model on the trend reversal reaction is slow, false signals are also fewer, give up before and after the part of the profit is also more. The disadvantage of this type of trading model is that it produces continuous losses during the consolidation market, which makes it unacceptable to investors. So the difficulty in designing trend following trading models does not lie in finding methods to capture trends, but in having a sound set of trend confirmation and filtering principles to avoid risks. In addition, the trend tracking trading model requires futures fund managers to hold positions for a longer period of time, usually more than 2-3 months, so futures fund managers are required to have a set of psychological control methods that are compatible with the trend tracking trading model.

3) Counter-trend based trading models

This type of model is a system that is based on the designer's statistics and then assumes that the market needs to be adjusted and builds a position to trade in the opposite direction. The difference between this and trend-based trading models is that trend-based trading models can be adjusted automatically, whereas counter-trend-based trading models must come with a set of stop-loss conditions due to the incalculable risk that often comes with operating in the opposite direction of the primary trend.

2, the basic analysis of the trading model

The basic analysis of the trading model refers to the trader to use the market outside the data and information, through the examination of all the information affecting the basic economic relations, and quantitative analysis of such factors, the establishment of a database, from which to judge the equilibrium price of the market for the investment model. The characteristics of the model are mainly: to provide a good analytical basis for large-scale capital into the market; strong theoretical basis, easy for the investment public to accept; for the short-term and timing is not very helpful; the difficulty of information collection; the analysis of lagging behind the market price; the analysis of the subjective nature of the strong.

The following is an introduction to the "value assessment" and "assessment points" two basic analysis of the trading model.

1) value assessment trading model

Futures prices on the spot price will produce mutual traction, according to statistics, in the past 10 years, China's soybean futures price and spot price correlation coefficient of 0.9. And for the futures prices generated by the futures market, the futures market participants, including the spot traders and speculators, the futures price of the same commodity has its own judgment, and due to the the vast majority of participants in the mature futures market are speculators, the volume of the futures market is often several times or tens of times the volume of spot trade, so the futures price is not only determined by the spot price and storage costs, in addition to the cost of pricing also includes capital pricing component. Therefore, as the basic analysis of the futures fund trading model, but also include the futures market speculative factors: futures prices = (spot prices + storage costs) × speculation coefficient. Speculation factor is determined based on contingencies, market speculative funds and other circumstances.

2) Integral assessment trading model

The main disadvantage of the basic analytical trading model is the difficulty of information collection caused by information asymmetry, analysis lags behind the market price and analysis of the strong subjectivity, but with the development of information technology and trading system, the fair **** enjoyment of the information will further narrow the asymmetry of the information, the acquisition of up-to-date information is also relatively easy, the difficulties are The difficulty is how to identify the authenticity of information, primary and secondary and overcome the influence of excessive subjective judgment in information processing. The main steps of the integral evaluation trading model are as follows:

A. Determine the analytical factors

In order to keep the analytical statistical factors comprehensive, the number of analytical factors in both long and short sides should not be too small, generally not less than 5. If the supply and demand analysis factors, to soybean futures, for example, supply and demand factors include: forecast planting area and actual planting area factors; forecast production and actual production factors; soybean import and export volume; soybean crushing and processing volume; inventory factors; contingency factors, and so on.

Again, the cyclical analysis of factors, but also to soybeans, for example, cyclical analysis of factors include: March-April around - the U.S. and China soybean planting period, planting area prediction factors, while the South American new soybeans began to go on the market, the price of the bottom. 5-8 Around May-August - the weather and production of soybeans in the United States and China as the main analytical forecast factors, the arrival of the peak season for consumption, the price of soybeans from the early slow rise to July, August soybeans by the impact of fluctuating factors such as the green and hot weather, the price of the year reached its peak. around September-November --Chinese and U.S. soybean actual harvest factors, South American soybean sowing area is expected to factors, after October due to the Chinese and U.S. new soybeans on the market, the price once again fell back to the lowest price area of the year.

B, to determine the analysis of the time period

No matter what kind of trading model of the analytical method, need enough statistical analysis of sample data, in order to ensure the reliability of the statistical results, and therefore to go through more than one cycle, such as agricultural products, growth cycle, the economic cycle of the metal, etc., which should contain unexpected events or political factors, in order to test the trading analytical model to cope with the ability and ability to control risk.

C. Determine the score value

The method of determining the score value can use the ordinary positive and negative score method, weight score percentage value method, etc., the score value of the positive factors for the positive value, the score value of the negative factors for the negative value, there is no clear positive and negative tendency of the factor is taken as 0 points.

D, calculate the score results

The score of the factors affecting the cumulative value of the score, the score results, the score is positive, the trend of the market is mainly up; score is negative, the trend of the market is mainly down; score 0 or close to 0 points, the market will be in consolidation.

E, score tracking system

The occurrence of different events and the passage of time changes, the impact of the factors on the price is not the same, such as emergencies on the price of the impact of the changes in the influence of the event will gradually fade, so the factors to be scored on the value of the continuous adjustment to determine the score results, adjusting the results of the decision-making on the trading model.

3, mathematical measurement trading model

Mathematical measurement trading model refers to the designer according to modern investment theory, a large number of statistical analysis of historical trading data, from which to find out a certain law, in the market deviation or a specific situation when investing in the model, such as the arbitrage trading model, the short jump trading model.

Categorized from the user's point of view, there are mainly the following two kinds: one is the analytical trading model, the other is the operational trading model, there is a considerable difference between the technical analysis trading model and the basic analysis of the trading model:

1, the analytical trading model focuses on foresight, for the analysis of the market trend in advance; and the operational trading model focuses on the Reactive, when the market has appeared a certain price should be taken trading decisions.

2, analytical trading model focuses on the individual benefits of a section of the market market requires high accuracy, ignoring the analysis of unfavorable market conditions; and operational trading model focuses on the overall benefits of the actual combat, the requirements of the trading model of the market for all the circumstances of the revenue generated by the results of the overall assessment.

3, the biggest difference between the two is that the actual operator has to face pressure from all sides, including the market, investors, fund managers themselves and other pressures, so the design of the model should also include how to control the factors of psychological pressure through a certain method, the effective implementation of the signals issued by the trading model.