AU2011381553A1

AU2011381553A1 - Smart system and method for dynamic strategies in statistical arbitrage trading

Info

Publication number: AU2011381553A1
Application number: AU2011381553A
Authority: AU
Inventors: Francis CHEOK
Original assignee: Individual
Current assignee: Individual
Priority date: 2011-11-16
Filing date: 2011-11-16
Publication date: 2013-09-12
Also published as: WO2013071330A1

Abstract

A smart computer-implemented system and method, wherein dynamic trading strategies for statistical arbitrage of financial instruments are automatically created and fine-tuned constantly according to the real-time market conditions. A completely new approach to the computation of co-reliance parameters not only simplifies the implementation of statistical arbitrage, but also solves the technical problems of a traditional model such as the uncertainty of the suitability of trading pairs, the scarcity of the trading opportunity and the static trading strategy in dynamic real-time market conditions. Not only have dynamic and smart trading strategies improved the profitability and efficiency of statistical arbitrage trading, but its practical simplification has also made it a hedging system to be readily applied in the financial industry. It is designated for a plurality of different financial products and trading instrument pairs.

Description

WO 2013/071330 PCT/AU2011/001475 SMART SYSTEM AND METHOD FOR DYNAMIC STRATEGIES IN STATISTICAL ARBITRAGE TRADING Copyright Statement [0001] All of the material in this patent document is subject to copyright protection under the copyright laws of Australia and other countries. The copyright owner has no objection to the facsimile reproduction by anyone of the patent document or the patent disclosure, as it appears in official governmental records but, otherwise, all other copyright whatsoever are reserved. Technical Field [00021 The present invention relates generally to a computer-implemented system and method, wherein dynamic trading strategies for statistical arbitrage of financial instruments are automatically created and fine-tuned constantly according to the real time market conditions. Background Art [0003] One of the renowned trading strategies is pairs trading, which is also referred to as statistical arbitrage. This strategy evaluates pairs of instruments that are known to be statistically correlated. Knowledge of this trend creates an opportunity for profit, as on the occasions when these trading instruments break correlation for an instant, the trader may buy one and sell the other at a premium. The concept and technique of pairs trading have been well recognised by the industry. As a market neutral trading strategy, also called long-short strategy, pairs trading theoretically enables traders to profit from any given market conditions: uptrend, downtrend or sideways movement. It is designed to exploit short-term deviations from a long-run equilibrium pricing relationship between two trading instruments. 1 WO 2013/071330 PCT/AU2011/001475 [0004] Although statistical arbitrage has minimal downside risk theoretically, practically its traditional model has three fundamental technical setbacks, which are discussed as below. Summary of invention Technical Problem [0005] 1. The uncertainty of the suitability of trading pairs. This uncertainty comes from two main sources: one is from the methodology in making the determination of the suitability of the pairs; the other is from the dynamics of the changing markets, which may prove that a good pair has been turned into a bad pair due to the change of the market condition (pertaining literature or research in this area is extremely limited). [0006] The mainstream methods currently applied in the determination of effectiveness of the pairs for a long-short strategy have sought to identify the presence of a statistically significant mean reverting relationship between the instruments traded. The strength of the mean reversion is usually measured by the correlation coefficient or the cointegration of the pair. Thus the correlation or cointegration becomes the obvious pair selecting rule. [0007] Correlation is a numeric measure of the level of co-dependence in a linear statistical relationship between two or more random variables; it is measured on returns rather than prices. Nevertheless, the calculation of correlation usually needs to perform a complex economic model (Luenberger, 2002). Due to its complexity in computation, correlation is therefore often not suitable to be used in real-time trading systems (Vischer, 2010). All researchers recognise that one of the most important aspects of pairs trading is the correlation; however, very few papers deal directly with the issue. Even the latest literatures published by Schizas et al. (July 2011), Cummins 2 WO 2013/071330 PCT/AU2011/001475 (February 2011, December 2010) and Dunis et al. (March 2010) did not reveal any data regarding the correlation. [0008] Cointegration is a quantitative technique based on finding long-term relations between asset prices introduced in a seminal paper by Engle and Granger (1987), for which they earned the Nobel Prize (2003) for statistics. Cointegration helps identify potentially related pairs of financial instruments. However, the difficult two-step procedure and sampling complication as indicated by Dunis et al. (2010) and Cummins (2010) have demonstrated that the cointegration approach, the same as the correlation methodology, is not practical to be applied in real-time trading systems. [0009] Furthermore, both correlation and cointegration approaches use fixed-term historical data (Schizas et al., 2011; Dunis et al., 2010; Cummins, 2011, 2010; Nath, 2003), which cannot detect on-going changes to the markets in the dynamic real-time trading environment; thus the validity of their previous stationary statistical relationship could evolve or undergo some form of structural change or may indeed break down entirely during the lead-up period. There is no evidence to support that the stationarity proved during the in-sample period has the guarantee to continue to be the same without a time limit (pertaining literature or research in this area is extremely limited). It becomes necessary to re-validate the correlation or cointegration from time to time when the system is running continuously without a time limit. Obviously, the impracticality of these statistical methodologies has to'be improved in order to make a continuously applicable trading system on a real-time basis in the real world. [0010] 2. The scarcity of trading opportunities. One of the most concerned conditions for pairs trading is the amount the pair correlated, as it requires a high correlation coefficient for security. The second prerequisite is the breaking of correlation. This is a paradox. The high correlation requirement is a technical dilemma for pairs trading, because the higher the correlation, the less likely that a breaking of 3 WO 2013/071330 PCT/AU2011/001475 correlation will occur; thus it will give very little chance to break its high correlation for traders to seize the opportunity. (00111 There is a difficulty in discovering the balance between correlation and deviation, where the right balance is an optimal point to open a trading position (pertaining literature or research in this area is extremely limited). How far away should the deviation be from its equilibrium mean to be sufficient to trigger an opening of a position without considering the degree of correlation? Without having done plenty of research and experiments, one cannot easily answer this question. However, it is even harder to answer when the question is asked with consideration of the degree of the correlation. If the trading system is designed to trade with a large deviation, it will be quite riskless but it will have little opportunities to trade. The same mechanism applies when the pair correlation is too high, it will have too few chances to achieve a large deviation. If the system is designed with too much protection with high correlation and large deviation, it might practically have no trading activities because these extreme situations will not occur regularly. Trading system developers must find a mathematical relationship between correlation and deviation for a long short strategy to optimise the tradability and profitability (pertaining literature or research in this area is extremely limited); too much emphasis on the correlation (or cointegration) cannot create a profitable trading system, because there are no meaningful trading activities generated. [0012]. 3. The static trading strategy in dynamic market conditions. No one has the evidence to support that the stationarity proved during the in-sample period has the guarantee to continue to be the same without a time limit in the real-time dynamic market conditions. In fact, to the contrary, through our long-term research, the static trading strategy used in the traditional statistical arbitrage is quite problematic; it has the direct responsibility for most of the losing trades. 4 WO 2013/071330 PCT/AU2011/001475 Solution to Problem [0013] The present invention not only has addressed the above setbacks of the traditional statistical arbitrage, but has also created a brand new approach to overcome these setbacks and limitations, and achieved the establishment of a profitable real-time trading system. [0014] 1. To conquer the uncertainty of the suitability of trading pairs, this invention utilises a very practical, simple and reliable method to gauge the correlation of the pairs; and this method is utilised to calculate a new parameter that is called the "co reliance rate", which as a new concept replaces the traditional concept of correlation. The higher the co-reliance rate the better suitability the pair will be. Therefore, the suitability for pairs trading will be easily determined by the value of the co-reliance rate, and the issue of uncertainty in suitability is solved by this new concept of co reliance rate. [0015] The more detailed calculation procedures for co-reliance rate are indicated clearly in the below embodiment Example 1. This great simplification not only provides a new parameter to measure the interdependency of the two trading instruments, but also offers many advantages when it is applied to a real-time trading system using the long-short strategy. [0016] 2. To deal with the scarcity of trading opportunity and achieve an optimal balance between correlation and deviation, and maintain good tradability and profitability, this invention has found a state-of-the-art solution. With long-term empirical data, this invention has struck a fine balance between correlation and deviation. However, its overall approach particularly in making a trading decision is completely different to the traditional practice. 5 WO 2013/071330 PCT/AU2011/001475 [0017] 3. To replace the static trading strategy with a dynamic trading strategy, the present invention has greatly improved not only its certainty of the suitability and its scarcity of trading opportunity, but also the profitability, which is paramount for the purpose of the whole trading system. As the correlation has been replaced by the co reliance, this invention has theoretically set the dynamic relationship between co reliance and deviation to be non-linear and inversely proportional. When the co reliance is high the deviation will be set low; when the co-reliance is low the deviation will be set high. This dynamic non-linear, inversely proportional relationship together with many other dynamic trading parameters have not only fundamentally changed the methodology of traditional pairs trading, but have also made it simpler and smarter in dealing with the real-world trading scenarios. Through presetting the lowest co-reliance rate allowed for pairs trading, the system will automatically detect and reject a pair for trading process when its co-reliance rate has reached the presetting limit during the real-time trading session. [00181 In practice, this invention actually solves the problems with great sophistication. It creates more new concepts, such as "anti-trade strength", "pro-trade strength", "pro-hedge strength" and "trend strength", in an attempt to balance multiple factors (not just the co-reliance and deviation) for their contribution in making comprehensive trading decisions. All detailed computation procedures are specified clearly in the below embodiments of Example 1 and Example 2 with in-depth discussions. Advantageous Effects of Invention (0019] The present invention has improved and changed the traditional statistical arbitrage in many areas. The advantageous effects of this invention will become better understood with reference to more detailed description, drawings and mathematical 6 WO 2013/071330 PCT/AU2011/001475 equations demonstrated under embodiments of Example 1 and Example 2. However, a brief summary of advantageous effects is presented as below: [0020 1. The co-reliance. Not only can the introduction of co-reliance solve the problem of the scarcity in trading opportunities and the uncertainty of the pairs trading suitability, but can also effectively establish a dynamic trading mechanism to create a smart and dynamic trading system. This is done by the constant checking of the evolving co-reliance; once the co-reliance rate decreases or increases, this new invented system will automatically detect these changes and increase or decrease the requirement of the price deviation respectively according to the change of co-reliance. This system will also automatically identify and discard a pair for trading process when its co-reliance rate has reached the presetting low limit during the real-time trading session; thus the system provides a protection from unsuitable trading with a dynamic trading strategy according to the constant varying market conditions. In contrast, the traditional model of statistical arbitrage fails to detect any changes of correlation during the real-time trading session; it cannot protect itself due to the incapability in assuring the pairing suitability when it moves with time. [0021] There are five advantages in replacing the correlation with co-reliance; these advantageous effects are analysed with more detail as follows: (0022] a) The "co-reliance rate" can be computed on a real-time basis continuously with any time intervals, thus it can form a dynamic semi-real-time moving average to reflect the on-going changes of the interdependency of two trading instruments. The traditional statistical correlation or integration approach can only produce static data to reflect the past, but fails to update itself with the current market situation. This revolutionary change has enabled the creation of an intelligent trading system that has the ability to constantly adjust and fine-tune its trading parameters and automatically make different trading decisions according to the real-time market conditions. 7 WO 2013/071330 PCT/AU2011/001475 [0023] b) This new parameter can ignore the test of stationarity of cointegration or correlation. There is no out-of-sample period for it; all real-time data collected will automatically become in-sample and will be used to update the market conditions. It is not static but dynamic and live data which moves with time. [0024] c) It is more intuitive and much easier to understand the interdependency by its method of calculation compared with the abstract concept of correlation and cointegration; particularly the "co-reliance rate" is measured by prices rather than returns. [0025] d) Its mathematical meaning and implication for interdependency are different to correlation and cointegration, but it is more than competent to replace them when it is applied to a real-time trading system using the long-short strategy. [0026] e) Its simplicity, practicality, reliability and intuitivity make it more adoptable for other future real-time applications. [0027] 2. The trend strength. By calculating all the ratios of current deviation to standard deviation of each trading instrument, the trend strength is easily obtained. This is a real-time live data computed every 5 or 10 seconds depending on how many trading instruments are incorporated in the system. The creation of this parameter facilitates this invention the capability to judge the current market condition on a real time basis; and then based on this capability the trading strategy can be refined accordingly and dynamically. There is no such a parameter or method available in the traditional pairs trading. [0028] 3. The pro-trade strength. This is an empirical formula that mainly measures the distance between two ratios of current deviation to standard deviation of the two trading instruments of the pair. It is a real-time live data that changes every time when the quoting prices are varied. Its value is directly proportional to the overall deviation of the two trading instruments, and the larger the value the higher the force to revert 8 WO 2013/071330 PCT/AU2011/001475 to the mean. Thus it determines the strength of mean reversion, which implies the probability of "winning" if it is traded. The pro-trade strength actually refers to the strength of winning. The bigger the pro-trade strength the more likely the system will trigger a trade. This approach is totally different to the traditional pairs trading strategy, where the traditional way uses the spread deviation to measure the reversion strength. However, the spread standard deviation and its mean are determined from the historical data in the in-sample period; therefore they are static and cannot be updated with time. The present invention has no such a problem; both the standard deviation and the mean are updated with moving time, which makes it a much more superior system with great difference to the accuracy of trading decisions. [0029] 4. The anti-trade strength. By definition, this parameter measures the strength of "losing" at the time of calculation. This empiricalformula collects all factors that may affect the profitability, and assigns them with proper weighting, making this formula the most comprehensive and sophisticated in this invention. The calculation of this real-time anti-trade strength involves the price change rate (similar to volatility), price range factor, co-reliance rate and trend strength. Only when the real-time pro-trade strength is larger than the dynamic anti-trade strength, the system will then trigger a trading decision to open long-short positions. The dynamic prerequisite for opening trade positions herewith enables the system to generate dynamic trading strategies, which change constantly with the real-time market conditions, hence making better trading decisions. Compared with the traditional pairs trading, this is much more advantageous because the traditional method does not take so many factors into consideration; it makes simple and primitive trading decisions once the pair spread has enough deviation, disregarding other factors such as volatility, correlation, price range and trend strength. The anti-trade strength is a significant step ahead of the traditional strategy, and it is one of the most important features to make the system consistently profitable. 9 WO 2013/071330 PCT/AU2011/001475 Brief Description of Drawings [0030] FIG. 1 is a flowchart of a general pairs trading system which is applicable to all hedgeable financial instruments. [0031] FIG. 2 is a flowchart depicting the detailed relationships and procedures between different components and parameters. [0032] FIG. 2A is a flowchart illustrating the relationships and procedures between different parameters, and revealing how a trading decision is concluded. [0033] FIG. 3 is a flowchart following FIG. 1, FIG. 2 and FIG. 2A, disclosing more details of mathematical relationships and procedures for making trading decisions. [0034] FIG. 4 is a flowchart following FIG. 1, illustrating a general hedging system which is applicable to all hedgeable financial instruments. [0035] FIG. 5 is a flowchart following FIG. 1 and FIG. 4, depicting all detailed relationships and procedures for the calculation of pro-hedge strength. [0036] FIG. 6 is a flowchart following FIG. 1, FIG. 2, FIG. 4 and FIG. 5, disclosing more details of mathematical relationships and procedures for making hedging decisions. [0037] FIG. 7 is a flowchart depicting the new strategy in protecting existing investment portfolios for hedge funds. Description of Embodiments [0038] The present invention will be embodied in a computer program without difficulties by those of ordinary skill in the art as the following description will disclose all the details. In the effort to explain all the elements, mathematical equations and flowchart figures are all set forth to provide a thorough understanding of the present 10 WO 2013/071330 PCT/AU2011/001475 invention. Nevertheless, it should be understood that illustrations and descriptions of specific embodiments are provided only as representative examples. With the spirit and scope of the present invention, all kinds of modifications, variations and improvements to the general applications herewith will be readily noticeable to those of ordinary skill in the art, and the general principles characterized herein can be harnessed to other embodiments and applications. Additionally, those areas in which it is believed that those of ordinary skill in the art are familiar have not been mentioned herein in an attempt to not unnecessarily obscure this invention. Therefore, the present invention is not to be limited to the embodiments demonstrated, but is to be accorded the widest possible scope consistent with the principles and features revealed herewith. Example 1 [0039] The embodiment of Example 1 herewith is the best contemplated mode of the present invention, it demonstrates how this smart system works in stock index trading. This invention is not limited to trading indices of all kinds; it can be applied for all kinds of derivatives and other tradable financial instruments, such as stocks, securities, commodities, futures, options, foreign exchange, bonds and the like, as long as the trading instruments are hedgeable. It is designated with the capacity to run in a computer system with unlimited different kinds of financial products and a plurality of financial instrument pairs. This embodiment is just showing one of the examples in practical applications. [0040] A general outline of acts corresponding to a normal embodiment of the inventive method is shown in FIG. 1, where step 101 is the first step to start selecting trading instruments. Let's assume there are 20 global stock indices in the futures market that we have selected as our trading instruments. They are the American Dow Jones Index, Nasdaq 100, S&P 500, Japanese Nikkei 225, Swiss 30 Index, Australia SPI 200, Italian 40 Index, Spanish 35 Index, British FTSE 100, Hong Kong Hang Seng 33 11 WO 2013/071330 PCT/AU2011/001475 Index, Swedish Index 30, Germany DAX 30, French CAC 40 and etc. Through analysis, we then move to step 102 of FIG. 1 to combine these 20 indices and make pairs; we assume to predetermine 100 index pairs for this system to monitor and trade. The selection and the combination of these financial instruments and pairs may be predetermined with or without knowledge of their historical characters, because this smart pairs trading system has the capability to automatically detect their suitability during the real-time trading period. The number of 100 pairs herein is just for this exemplary system. The present invention is designed with the capacity for taking unlimited number of pairs as long as the computing resources are adequate to support it. [0041] Some of the predetermined financial instrument pairs can be eliminated from the current system if their co-reliance rate is consistently lower than a preset value, which is the lowest level of co-reliance rate allowed to run in the system; this elimination will release more computing resources for other more suitable pairs to join the system in an attempt to improve the system's efficiency. [0042] Now we move to step 103 of FIG. 1 to create a pairs trading timetable; because we need to make sure our database has detailed information about the trading hours for each individual said index as each global index operates in different time zones. We then create another trading timetable to stipulate the time that both indices of each pair are tradable within this stipulated time frame. Step 103 may not be needed if all the trading instruments are in the same trading hours. [0043] The real-time trading data 111 is an external source that we need to set up a real-time data collector 133 to connect and receive data feeding through the Internet or other means. Our database 134 also requires to obtain feeding of external historical data 112 by downloading through the Internet or other means, which usually is a one off process needed only at the beginning in setting up the system. All historical data of each trading instrument must include a maximum price, a minimum price and a 12 WO 2013/071330 PCT/AU2011/001475 closing price for each trading day. All the prices in the system are the average of a bid price and an ask price. After connecting the real-time data collector 133 with the database 134, our system will have the continuous flow of real-time trading data with the support of historical data, and our database 134 will then have the capacity to provide technical parameters for each global index or pair. [0044] The heart of this arbitrage system is the analytical module 135, which does a wide range of calculations for making trading decisions; this will be explained in-depth later. This example, as depicted in FIG. 1, shows that once the system makes a trading decision, it has two remaining steps left. One is sending the trade signals to a trading module for automatic execution as illustrated by step 136. Another is transmitting the trade signals to external traders through the internet as illustrated by step 137, where a plurality of external traders can share the technology and benefit from it. [0045] How the analytical module 135 functions is illustrated in detail by FIG. 2. To understand the whole procedures step by step, below we show how these parameters for each single index instrument are defined and named in this system. [0046] Let's make V stand for Price Change Rate Of A Day, then S100. Day Maximum Price - Day Minimum Price Close Price of Previous Day [0047] Price change rate 201, as shown in FIG. 2, is a simplifying method to calculate the price volatility; it directly reflects the movement of price volatility although it is not exactly the same as the price volatility. The V here can be replaced by the real calculation of volatility, providing that this V value can be updated at each sampling interval. [0048] This current embodiment Example 1 uses a daily interval for sampling calculation; but this interval can be altered to a higher frequency sampling model 13 WO 2013/071330 PCT/AU2011/001475 without any fundamental change; i.e. this system can easily change its sampling interval without any difficulties. High-frequency or low-frequency intervals do not make any fundamental difference, but usually a higher frequency produces a more detailed insight, which enables the system to make better trading decisions. However, a higher frequency interval also requires many more computing resources, which might slow down the whole system or even make it unworkable depending on the hardware and other software capacity. [0049] The daily interval applied in this embodiment is called the low frequency of a predetermined long time interval, which is usually predetermined to be longer than or equal to 24-hours, or any other time such as 8-hours, 4-hours, 2-hours or 1-hour; this interval is mainly used as the frequency for important historical data processing, calculating, updating and finally storing into the database. Another frequency of data process is called the high frequency of a predetermined short time interval, which is usually predetermined to be longer than or equal to 5-seconds, or any other time such as 60-seconds, 40-seconds, 20-seconds, 10-seconds or 3-seconds; this interval is mainly used as the frequency for calculating, updating and saving of dynamic real-time data and parameters. [00501 Let's make V 252 stand for Average Daily Price Change Rate Over 252 Days, then V2s2 (Vk) [Math. 2] k=1 [00511 where N = 252, Vk is the price change rate of day k. V 252 is actually an annual moving average of daily price change rate (202, shown in FIG. 2) as it is updated every day to obtain a new average of the latest 252 days, i.e. a new day will be integrated and the oldest day will be eliminated at the same time. Thus V 252 is refreshed and varied on a daily basis. If a higher frequency sampling interval, such as 8-hours or 14 WO 2013/071330 PCT/AU2011/001475 other hours, is used here, then V 252 will be updated on every sampling interval. The reason it is "annual" is because there are approximately 252 trading days in a year. [00521 Let's make V 2 3 stand for Average Daily Price Change Rate Over 23 Days, then N

V

2 3 oVk) [Math. 3] k=1 [0053] where N = 23, Vk is the price change rate of day k. V 23 is also a moving average with the same mechanism as V 252 . We may also call V 23 a monthly moving average of daily price change rate (203, shown in FIG. 2), because there are roughly 23 trading days in a month. The obvious purpose to have these two annual and monthly moving averages is to gauge the long-term and short-term effects on price volatility. These two terms can be varied to longer or shorter times. In this embodiment example, "long-term" refers to a fixed term of one year, unless it is set for other predetermined periods; and "short-term" refers to a fixed term of one month, unless it is set for other predetermined periods. [0054] Let's make A 252 stand for Price Range Factor Over 252 Days, then Period Maximum Price - Period Minimum Price

A

252 = 100 (Period -Period Minimum Price [Math. 4] [0055] A 252 is not a moving average, but its calculation includes the day of that moment it is calculated, thus A 252 is also a data that changes with time, updated on a daily basis. This is another measurement of the price volatility, and its period can be varied to longer or shorter times; because in this example we use an annual period, we can also call A 252 a long-term price range factor. [0056] Let's make S stand for Standard Deviation Over 23 Days, then 15 WO 2013/071330 PCT/AU2011/001475 N 2 S = N-1 X) ,[Math.5] k=1. [0057] where N=23, Xkis the close price of day k, and X is the mean of close prices over 23 days. S is not a static standard deviation; it is updated on-a daily basis or on every data sampling interval, working with the similar mechanism as the V 23 moving average. The entire sampling period (not to be confused with the sampling interval) of 23 days is actually just an example here. This invention allows any other -sampling periods selected for the calculation of standard deviation because this period, such as 60 days, 30 days or 15 days, depends on the frequency of sampling intervals and many other factors. Because in this example we use a monthly period, we can also call Sa short-term standard deviation. [0058] Let's make D, stand for Current Price Deviation (or Current Deviation), then D= Current Real-time Price - X [Math. 6] [0059] Let's make R stand for Ratio Of Current Deviation To Standard Deviation, then Dc R = [Math. 7] S [0060] Because in this example we use a monthly period to calculate the standard deviation, we can then also call R a short-term ratio of current deviation to standard deviation. The ratio R is a real-time live data changes with its current price. [0061] Above formulas are all for the single stock index. Below equations refer to the pair; they define the co-reliance relationship of the two indices. It is not an easy task 16 WO 2013/071330 PCT/AU2011/001475 to understand their co-reliance relationship mathematically. Below variables only refer to one day. Let's make: [00621 MAX1 stand for the maximum price of index 1 of the day; [0063] MAX2 stand for the maximum price of index 2 of the day; [0064] CLS1 stand for the close price of index 1 of the day; [0065] CLS2 stand for the close price of index 2 of the day; [0066] MINI stand for the minimum price of index 1 of the day; [0067] MIN2 stand for the minimum price of index 2 of the day; [0068] CLOSE 1 stand for the close price of the previous day of index 1; [0069] CLOSE2 stand for the close price of the previous day of index 2; [0070] Ca stand for co-reliance-max-rate; [0071] Cb stand for co-reliance-min- rate; and [0072] Ce stand for co-reliance-close-rate. [00731 If IMAX1-CLOSE1| IMAX2-CLOSE2| CLOSE1 CLOSE2 / MAX2 - CLOSE2 MAX1 - CLOSE1 Ca = CLOSE[Math.8] [0074] If IMAX2-CLOSE21 > IMAX1-CLOSE1I then CLOSE2 CLOSE MAX1 - CLOSE MAX2 - CLOSE2 Ca =)[Math. 9] CLOSE1 CLOSE2 [0075] If IMIN1-CLOSE1I IMIN2-CLOSE2| then CLOSE1 CLOSE2 ( MIN2 - CLOSE2, (MINI - CLOSE 1 Cb = k CLOSE2 \ CLOSE1 [Math.10] .17 WO 2013/071330 PCT/AU2011/001475 [00761 If IMIN2-CLOSE2| IMIN1-CLOSE1| CLOSE2. CLOSE1 ' then C - (MIN1 - CLOSE (MIN2 - CLOSE2 ) C= CLOSE1 - CLOSE2 [Math.11] [0077] If ICLS1-CLOSE1I ICLS2-CLOSE2\ then CLOSE1 CLOSE2 cc, (CLS2 - CLOSE2 (CLS1 - CLOSE1 [Math. 12] [0078] If |CLS2-CLOSE2| | CLS1-CLOSE1| CLOSE2 CLOSE (CLS1 - CLOSE1 CLS2 - CLOSE2 C CLOSED CLOSE2 [Math. 13] [0079] Finally, let's make C stand for Co-reliance Rate Of A Day; we can then conclude: C = (Ca + Cb + Cc) [Math. 14] 3 [0080] Above formulas show that co-reliance is a quantifying measure of the degree of interdependency in a linear non-statistical relationship of two random trading instruments; and it is calculated on prices rather than returns. These formulas and methodologies developed by this invention are particularly useful for the study of the interdependency of two trading instruments in the financial markets. The term co reliance rate herewith is not the same as previously published Pearson Correlation Coefficient or other correlation theory of statistics; its concept and definition are defined mathematically here according to the above methodology. A single co-reliance rate has no statistical implication; however a plurality of co-reliance rates can be interpreted in many ways to many useful statistical meanings. 18 WO 2013/071330 PCT/AU2011/001475 [0081] When the above co-reliance rate C > 0, it means the two financial instruments are proportionally correlated, i.e. they follow the same trend at a proportional degree. If one goes up, the other also goes up. Multiplying the C value by 100, the new result is the percentage of their co-reliance; and the co-reliance percentage will never exceed 100%; i.e. the co-reliance rate will never exceed 1. The higher the C value the more they are correlated in the same direction. [0082] When the above co-reliance rate C =0, it means the two financial instruments have no correlation. [0083] When the above co-reliance rate C < 0, it means the two financial instruments are inversely proportional in correlation, i.e. they split into two opposite trends at a proportional degree. If one goes up, the other goes down. The higher the absolute C value the more they separate into two opposite directions; but the negative co reliance rate never exceeds -1; i.e. the negative co-reliance percentage never exceeds -100%. [0084] Let's make C 252 stand for Average Pair Co-reliance Rate Over 252 Days, then N C2s2 (Ck) [Math. 15] NI k=1 [0085] where N = 252, Cis the Co-reliance Rate of-day k. C 252 is an annual moving average working with the same mechanism as the V 25 2. [0086] Let's make C 23 stand for Average Pair Co-reliance Rate Over 23 Days, then N C23 =(C) [Math. 16] k=1 19 WO 2013/071330 PCT/AU2011/001475 [0087] where N = 23, Ckis the Co-reliance Rate of day k. C 23 is a monthly moving average working with the same mechanism as the V 23 .The obvious purpose to have these two moving averages C 252 and C 23 is to gauge the long-term and short-term effects on price correlation. These two terms can be varied to longer or shorter times. [0088] One of the hardest technical problems in a statistical system is the appropriate size of the time range for data analysis and evaluation. If the selected time range is too large, it will take too many risks in being trapped into data mining (Chadwick, 2011); however, if the selected time range is too small, it will then take another big risk in being locked into a temporary data pattern, which will lead you into a disaster of misinterpretation of data. To avoid data mining, we choose to apply the monthly moving average; and to avoid being locked into a temporary pattern, we choose to utilise the annual moving average. The combination of a relatively long time range and a relatively short time range creates a well balanced mathematical model. This solution fits the current Example 1 exceptionally well, but other time ranges might fit better for other examples. This invention is not restricted to any other time ranges as long as they fit the actual applications. [0089] With the above known parameters, we create three brand new trading indicators - the Anti-Trade Strength for a pair, the Pro-Trade Strength for a pair and the overall current market Trend Strength. They come from the current market parameters that indicate the current market condition for the pair or market that we focus on. Their computation formulas are as below: [0090] Let's make Strend stand for the Current Overall Market Trend Strength, then N Strend = (R.) [Math. 17] k=1 20 WO 2013/071330 PCT/AU2011/001475 [0091] where N =20 (or any number of indices that we would like to focus on), Rk is the real-time ratio of current deviation to standard deviation of index k. Trend is a real time live data, which will be updated every 5 seconds in the system in an attempt to constantly grasp the current market condition. [0092] Let's make Santi stand for Anti-Trade Strength for a pair, then San= Wv+ Wc+ WA+ WfiV W7 S [Math. 18] [0093] where [0094] Wiv is the weighting of both of the long-term and short-term moving averages of price change rate of instrument 1 in a pair; [0095] Wc is the weighting of both of the long-term and short-term moving averages of co-reliance rate of a pair; [0096] WAis the weighting of the long-term price range factors of both instrument 1 and instrument 2 in a pair; [00971 W 1 v is the weighting of both of the long-term and short-term moving averages of price change rate of instrument 2 in a pair; and [0098] W-s is the weighting of the current market trend strength. Wiv = 6(Vi252 - a) (100|Vi2s2 - aj)P + g(Vi23 - y) (100|Vi23 - y|)" [Math. 19] [0099] where Vi2 52 and Vi 23 are respectively the long-term and the short-term moving averages of price change rate of instrument 1 of a pair; 6, 0, a, F, A and y are all positive empirical constant factors. The values of these constant factors in this Example 1 have been set to be: 6 = 13; 0 = 0.48; a = 0.8; k= 18; A = 0.59; and y = 0.85. 21 WO 2013/071330 PCT/AU2011/001475 Wc < (( - 100C252)|( - 10OC2s2|1 + &(p - 100C23)|p - 100C23|' [Math. 20] [0100] where C 23 and C 252 are respectively the short-term and the long-term moving averages of co-reliance rate of a pair; A, p, q,, ( and p are all positive empirical constant factors. The values of these constant factors in this Example 1 have been set to be: A = 0.006; p = 70; c = 2.48; # = 0.005; (= 69; and p = 2.33. WA = (Ai252 - T) |Ai2s2 - TIE + (Aii252 - T) IAii252 - TI E [Math. 21] [01011 where Ai2 52 and Ai2 52 are respectively the long-term price range factors of instrument 1 and instrument 2 of a pair; T and E are all positive empirical constant factors, which have been set to be: T = 10; and E = 0.5. Wiiv = (Vii252 - ) (100|VH2s2 - aJ)P + k(V 23 - y) (100|23 - y|)A [Math. 22] [0102] where Vii2s 2 and V 1 23 are respectively the long-term and the short-term moving averages of price change rate of instrument 2 of a pair; 6, 0, a, F, A and V are all positive empirical constant factors; which have already been specified in Math. 19. WTS = 1 (|Strena j - X) I(IStrend I - a)|I [Math. 23] [0103] where a, ) and 0 are all positive empirical constant factors, which have been set to be: a = 0.5; C = 383; and E = 0.5. [0104] Sant is a dynamic real-time data similar to Strend. Although these formulas look complex, the computation of the value of Santi takes less than a second thanks to modern computing technologies. 22 WO 2013/071330 PCT/AU2011/001475 [0105] Let's make Spro stand for Pro-Trade Strength for a pair, then Spro = i |Ri - Rii|I* - A [Math. 24] [0106] where R, and R 1 are the ratios of current price deviation to standard deviation for the instrument 1 and instrument 2 of a pair respectively, they are dynamic real time data; n, A and $j are empirical constant factors, which have been set to be: q = 0.386; A = 100; and @) = 5.92. [0107] All of the above empirical formulas and their constant factors are the results of long-term research; values of these factors all come from trial and error with long term observation and fine-tuning; however, variation of these values is allowed as long as they function for the purpose. Let's just take q, A and $ in Math. 24 as an example, if we increase the value of q and at the same time decrease the value of if, we might achieve a same final result of Spro; if we increase the value of $ and also increase the value of A, again we might obtain a same final result of Spro. Therefore, it is obvious that all aforementioned values of constant factors in above mathematical equations are all changeable, although they may also be fixed constants. [0108] From above detailed descriptions of all mathematical equations, it will be apparent to easily understand the flowchart of FIG. 2. You might have noticed that the same step number with broken lines are used in the previous figure - the same step number from previous figures means that this particular step or component or act is the identical one indicated previously, unless the drawings used a new step number. The step with broken lines is not a new step. Therefore, we may not explain in detail for those repeated steps because they have already been discussed previously. Let's look at FIG. 2 and continue our discussion. [0109] With all historical data and real-time data, the database 134 supplies all required data to compute all parameters, such as the daily price change rate 201, the 23 WO 2013/071330 PCT/AU2011/001475 annual moving average of daily price change rate 202, the monthly moving average of daily price change rate 203, the daily co-reliance rate 211, the annual moving average of daily co-reliance rate 212 and the monthly moving average of daily co-reliance rate 213. You may notice that the results of 201 and 211 will be sent back to the database 134 for storage as historical data for future usage. The calculation of the price range factor 280 and the ratio of current price deviation to standard deviation 232, as you may be aware, requires a direct passage of data from the real-time data collector 133. This requirement especially makes the ratio 232 a live data as it will vary each time it is recalculated. The live ratio.232 then passes on its value for the current overall market trend strength 233, thus making the trend strength 233 a real-time data too, changing every second. Apparently, all of the above steps require the usage of mathematical equations ranging from Math. 1 to Math. 17; but the computation should be very straightforward for implementation. [0110] Step 240 has six arrows pointing to it, as it is illustrated in FIG. 2 showing weightings of co-reliance parameters and other various dynamic trading parameters. The weightings herewith are the total of individual weightings of each trading parameter, in which their sources have been indicated by the arrow line. The total result of step 240 is then passed on to step 252, as it is depicted in FIG. 2A. It is apparent that the anti-trade strength 252 demands most of the resources; the contribution of each dynamic trading parameter for the anti-trade strength is calculated by each weighting formula ranging from Math. 18 to Math. 23, involving the annual moving average of co-reliance rate, the monthly moving average of co reliance rate, the annual moving average of price change rate, the monthly moving average of price change rate, the price range factor and the market trend strength. There are two real-time live parameters 280 and 233 as part of the final result of 252; this certainly makes value 252 also a real-time live data. [0111] In FIG. 2A, the only source for the calculation of the pro-trade strength 251 is the live ratio 232; it makes the value 251 also a real-time live data. In fact, the pro 24 WO 2013/071330 PCT/AU2011/001475 trade strength 251 is the most active real-time data in the system because it involves both real-time prices of instrument i and instrument ii. Naturally, in comparing the pro-trade strength 251 with the anti-trade strength 252, the result for step 260 will fluctuate in every moment. If Spro > Santi then a decision to open trades (step 271) will be activated, otherwise no action will be taken as step 272 requires nothing. [0112] Step 260 is the heart of this invention; it is the smart and dynamic trading strategy. As the trading prerequisite, step 260 is created through the comparison of two dynamic and comprehensive parameters, which include many other trading factors that reflect the real-time market conditions. Therefore, this smart trading system automatically reacts differently with various criterion to trade under the ever changing real-time market conditions. [0113] From here we have already established a smart trading system, which scans the market every 5 seconds to test any pairs that may fit the trading prerequisite. The scanning interval as 5 seconds is only a guideline, and this setting can be varied according to the quality of your computing resources. As long as the real-time prices of the pair fit to the below prerequisite, this system will activate the opening of a long short trading position: Pro- Trade Strength Spro > Anti- Trade Strength Sant [0114] The reason we describe this as a smart system is because it constantly adjusts and fine-tunes the trading condition automatically according to the market situation. For example, if the volatility has an obvious shift for some time, it will cause the Santi value to make a non-linear proportional shift accordingly. If the Santi value becomes higher, then it will require a greater Spro value to meet the trading requirement, thus it becomes harder to trade, vice versa. 25 WO 2013/071330 PCT/AU2011/001475 [0115] As the calculation of Santi value involves many factors, such as real-time trading prices, semi-real-time long-term volatility, short-term volatility, long-term correlation, short-term correlation and others, the result of this calculation is moving with the time according to the changes of these factors. Thus, any changes of the market condition reflected by these factors will be forwarded and included in the result of the Santi value, hence the trading condition or the trading strategy is dynamically modified accordingly. [0116] FIG. 3 is a flowchart following FIG. 1 and FIG. 2, disclosing more details of mathematical relationships and procedures for making trading decisions. We have assumed that financial instrument i and ii are the two instruments in the pair. Again in FIG. 3, the ratio of current price deviation to standard deviation 232, which has already been displayed in Fig. 2, will pass on values of R, and R 1 1 to step 363 for making comparisons; the results of these comparisons lead us to conclude which instrument should be bought and which instrument should be sold, as indicated by step 364 and 365. The instrument that possesses a higher value of ratio R (step 232) should be sold, because this is the one that has been overvalued and mispriced, and has broken its normal equilibrium mean with high probability of mean reversion. [0117] One of the subtle factors is the trend strength Strend, which is a market trending indicator. Normally its value should be around zero to indicate a peaceful market condition; when Strend = 0, it means the overall market is on a perfect sideways movement. When Strend > 0, the larger the value the more the overall market is on an uptrend. If Strend < 0, the larger the absolute value the more the overall market is on a downtrend. Thus one can easily judge the whole global market condition and its degree of trending with this quantified value of Strend [0118] The trend strength Strend is worthy of more examination here. Different trading instruments appear to be irrelevant to each other as long as they are not in the same pair; however, the computation of the Strend value makes them relevant. For example, when the financial crisis occurs, all the other irrelevant market prices plummet; and 26 WO 2013/071330 PCT/AU2011/001475 this extraordinary phenomenon will be picked up by the Strend value. Because the Strend value is part of the Sant- value, the alarm of the crisis will be passed on to the trading decision, alerting the system to create more stringent trading rules to avoid losses. This intelligent system design is obviously much more superior to any possible design based on the traditional long-short strategy, as it can pick up "trivial" irrelevant signals to safeguard the system, making it the best shockproof trading system available. This technique is not exclusively applicable for pairs trading - other hedging systems with other strategies or methods can utilise this invention as well. Example 2 [0119] The embodiment of Example 2 is one of the best contemplated modes of the present invention, it demonstrates how this smart system works for a hedge fund to protect an existing investment portfolio. This invention is not limited to hedging all kinds of stock shares; it can be applied for all kinds of derivatives and other tradable financial instruments, such as indices, securities, commodities, futures, options, foreign exchange, bonds and the like, as long as the trading instruments are hedgeable. It is designated with the capacity to run in a computer system with unlimited different kinds of financial products and a plurality of financial instrument pairs. This embodiment is just showing one of the examples in practical applications. [0120] A general outline of acts corresponding to a normal embodiment of this inventive method is shown in FIG. 4, where step 401 is the first step to start selecting hedging instruments and instruments needed to be hedged. Let's assume there are certain amounts of company shares in the existing portfolio; these shares have been diversified to six blue chip companies such as Wal-Mart, Boeing, Citigroup, Exxon Mobil, Microsoft and General Electric; and here they are considered as instruments needed to be hedged; these instruments will be referred to as hedged instruments in the following discussions. 27 WO 2013/071330 PCT/AU2011/001475 [0121] When the share market goes up, it is best to do nothing. When the market turbulence comes, the hedge fund has three choices: do nothing, hedge the shares or sell them. Suppose the hedge fund decides to protect its investment portfolio by hedging, then the present invention will be a great tool for this purpose. [0122] A good hedging not only can protect the existing investments, but can also bring profits. However, it is well-known that a bad hedging could bring disaster - not only is it unable to protect the existing investments, but it could also lose more money. Therefore, correct hedging needs to be done with correct methodology. [0123] First, we search for possible instruments which are suitable candidates for the hedging purpose. For example, possible instruments for hedging the Wal-Mart shares may be all retail stocks, retail sector index, the Dow Jones Index and the S&P 500, etc. Other possible hedging instruments for Citigroup shares may be all financial stocks, financial sector index, the Dow Jones and the S&P 500, etc. After analysing all possible instruments for hedging the above-mentioned six blue chip shares, let's assume that for each blue chip share we end up having a different group of suitable hedging instruments to match with; and each different group consists of 15 instruments. [0124] Now we have finished step 401 for selecting hedging instruments and instruments needed to be hedged. We then move to step 402 of FIG. 4 to combine these different groups of suitable hedging instruments to make pairs with each matching blue chip share, which are defined as hedged instruments. Therefore, we now have formed total 90 pairs for the hedging purpose, because each blue chip share can be hedged by 15 instruments; and these 15 instruments will be referred to as the hedging instruments in the following discussions. In order not to be confused in our later discussions, let's make the first instrument in the pair -instrument i (or instrument 1) be the hedging instrument; and the another instrument in the pair instrument ii (or instrument 2) be the hedged instrument. 28 WO 2013/071330 PCT/AU2011/001475 [0125] The real-time trading data 111 is an external source, for which we need to establish a real-time data collector 133 to connect to and receive from. Our database 134 also requires feeding of external historical data 112, which usually is a one-off process needed only at the beginning in setting up the system. After connecting the real-time data collector 133 with the database 134, our system will have the continuous flow of real-time trading data with the support of historical data, and our database 134 will then have the capacity to provide technical parameters for each instrument or pair. The heart of this hedging system is the analytical module 435 for pro-hedge strength calculation, which will be explained in FIG. 5. This example, as depicted in FIG. 4, shows that after the request to run the analytical module 435, the system makes a ranking list 436 with values of pro-hedge strength. [0126] You might have noticed that most of the drawings in Example 2 apply the same step numbers which are used in Example 1 - when this occurs in Example 2, it means that this particular step or component or act is identical to the one in Example 1. Repeated steps are depicted with broken lines; new steps have solid lines with fresh step numbers. Through observing drawings in Example 2, we see only a small number of steps are new; we may easily understand from here that Example 2 is a further application based on Example 1. Therefore, we need not explain in detail for those repeated steps as they have already been discussed. [0127] How the analytical module 435 functions is illustrated in detail by FIG. 5. To understand the whole procedure step by step, please refer to FIG. 1, FIG. 2, FIG. 2A, FIG. 4 and FIG. 5 as this example of embodiment has a very similar process to Example 1. By making some minor modification from our smart pairs trading system used in Example 1, the trading system becomes a hedging system, which is especially capable of providing best solutions and choices for hedging. We simply transform the trading trigger in the trading system to a hedging indicator by listing all the Pro-Hedge Strength (step 590, as it is shown in FIG. 5), which will be calculated as below: 29 WO 2013/071330 PCT/AU2011/001475 Pro-Hedge Strength Shedge = Spro - Santi [Math. 25] [0128] The pro-hedge strength is a measurement of hedging suitability for a hedging instrument to hedge against an existing investment. The higher the strength the more suitable it will be. The ideal scenario is when the hedging instrument has broken its long-term correlation with the hedged instrument and surged to the point it has become so far from its mean, creating a huge mean reversion force. At this point in time, the pro-hedge strength will be very high, and it is a hardly sought opportunity to sell and hedge if you want to hedge your investment; because the obviously overvalued and .peaked hedging instrument will soon revert to its mean. Selling this kind of hedging instrument to hedge is the safest decision; however this is only an ideal scenario. The present invention is a very useful tool in searching, identifying and creating the ideal scenario like this or close to this for fund managers to protect their investment portfolios. At any time as indicated by step 591, fund managers can always run this hedging system and find out what are the best available hedging instruments, how good they are and how the pro-hedge strengths are. With this hedging system, no more guesswork will be needed or done, and no more hedging is based on luck. [01291 It is important to note that the anti-trade strength in Example 2 is still the same step 252 in Example 1, but the pro-trade strength 551 in FIG. 5 is different to step 251 in Example 1; this is because either instrument of the pair in Example 1 can be bought or sold, but in Example 2 the instrument i, which is assumed to be the hedging instrument, can only be sold due to the fact that the existing invested instrument ii has already been bought. Therefore, the ri value of the formula Math. 24 for the hedging system changes to negative (q = -0.386) when Ra is larger than RI; while the rl value at step 251 is always positive (q = 0.386). The q value of Math. 24 under the hedging system has two values; when R, is smaller than or equal to R 1 , rj = 0.386; while Ril is larger than R, then r = -0.386. This special situation is illustrated with more details in FIG. 6. 30 WO 2013/071330 PCT/AU2011/001475 [0130] Step 610 uses data from step 232 to determine whether R 1 1 is larger than R 1 . If the answer is "Yes", step 622 indicates that S,. will be a negative value because n = 0.386, then step 623 shows an apparent result that S%,eg, will be a big negative value as Spr, is negative. Therefore, the conclusion arrives at step 624 that the instrument i is not suitable for hedging because Shedge is far too low; it is too risky. If the answer is "No", step 632 indicates that S,. may be a positive value because n = 0.386, then step 633 illustrates that Smet. may become a positive value. Therefore, step 634 demonstrates that the instrument I could be possibly suitable for hedging depending on its suitability ranking shown at step 591. [0131] This invention demonstrates that instrument i is not suitable for the hedging purpose when R, < R 1 1 , because the price of instrument i has already reduced in proportion to more than the existing investment's price (instrument Ii), which implies that the price of instrument i has much less probability to further decrease compared with those that have been overpriced with mean reversion force. [0132] Hence, this new hedging system will list all 15 SI'ed values of pro-hedge strength for each blue chip share which the hedge fund holds. Now we have a hedging system showing us 15 possible hedging instruments with clearly quantified values that directly indicate their ranking of suitability for hedging from the best to the worst. It becomes a simple procedure from a complex issue, when all you do is just type in the name of the hedged instrument and press a button. [0133] We might choose the two top ranking instruments with the highest pro-hedge strength as our choices to hedge against one blue chip share. The reason in choosing two to hedge one is to diversify the risks; certainly, discretion to choose more or less is exercisable as long as this fits one's needs. [0134) With this hedging tool, you can easily find 12 best hedging instruments or more to hedge against your 6 blue chip shares at any time you need. These 12 choices 31 WO 2013/071330 PCT/AU2011/001475 come from a very sophisticated system, which evaluates all the possibly available factors on a real-time basis with a comprehensive algorithm. The results of the pro hedge strength will vary each time you run it, because its calculation is always based on the most current market conditions and dynamic real-time data. As we have shown in Example 1, this is an intelligent hedging system that automatically refines its hedging strategies according to the variation of the market conditions. We can preset an optimal boundary - a set value of the pro-hedge strength - which will be considered the best opportunity to hedge. When the hedging system automatically detects that the pro-hedge strength passes this boundary, it will trigger an alert for users, making audio signals at the same time displaying a visual notice showing the names of the hedging instrument and the hedged instrument together with the value of the pro-hedge strength and time. [0135] To further exploit its advantageous effects, we show another strategy as a reference here, as depicted in FIG. 7. Let's just focus on one blue chip share at a time as an example. Suppose we have run the present hedging system to obtain a hedging suitability ranking list. as illustrated by step 701 in FIG. 7; we then select the two hedging instruments from the top of suitability ranking list as shown in step 702; as an example, we assume these two hedging instruments are Nasdaq 100 index and Intel shares. Following step 703, we sell the Nasdaq 100 index and the Intel shares to hedge against the existing shares of Microsoft, which have already been assuredly bought in as an investment. [0136] After some time, we rerun the present hedging system (step 710) to re-check their pro-hedge strength to see the strength of mean reversion. In order to determine the market trending, we calculate the current overall market trend strength (step 711) to evaluate whether it is in a downtrend or uptrend (step 713). If the market does not continue to go down, we will then execute step 750 to release our hedging in an attempt to let our investment profit. However, if the market does continue to go down (step 713), it will be very likely that the Nasdaq 100 index and Intel no longer have the 32 WO 2013/071330 PCT/AU2011/001475 highest pro-hedge strength, and they might have been replaced by the Dow Jones index and IBM as the top ranking of the best hedging choices (just as an example, step 714). With this information, we first buy back and close the Nasdaq 100 index and Intel shares taking the profits that both instruments have generated (step 750); then we sell Dow Jones and IBM to re-hedge the existing shares of Microsoft (step 715). After some time, again as FIG. 7 illustrated, we go back to rerun this hedging system (step 710) and repeat the previous process. If the market still continues to go down, we then take the profits and find other better hedging choices again; we can continuously keep using and profiting from this strategy of hedging until the market changes and goes up. [0137] Above example demonstrates the capability of this smart system in an attempt to protect an investment portfolio of a hedge fund during a market downturn period. Not only can this smart system protect the existing investments, but it can also profit even during a market downturn period. The most important feature of this invention is its capability to constantly refine its strategy automatically in accordance with the rapidly changing global markets. (0138] Although the present invention has been described in considerable detail with reference to certain preferred versions thereof, other versions are possible. Therefore, the spirit and scope of the appended claims should not be limited to the description of the preferred versions contained therein. Citation List Patent Literature [0139] Chadwick, S., US Patent 7,921,050 B1, System and Method for Analyzing Data Associated with Statistical Arbitrage - Filed October 30, 2007. Date of Patent April 5, 2011. Column 1, lines 58 to 68. 33 WO 2013/071330 PCT/AU2011/001475 [01401 Vischer, A., S. Reib, US Patent 7,788,166 B2, Implied Index Correlation and Dispersion - Filed May 23, 2006. Date of Patent August 31, 2010. Column 1, lines 20 to 33. Non Patent Literature [0141] Cummins, M., December 2010, Empirical Analysis of Optimal Statistical Arbitrage Trading on the Irish Stock Exchange, http://www.mark-cummins.com/ resources/ MCOptimStatArbISEQ2.pdf. [0142] Cummins, M., February 2011, Optimal Statistical Arbitrage: A Model Specification Analysis on FTSE 100 Stocks, http://www.mark-cummins.com/ resources/MC_ OptimStatArbFTSE100.pdf. [0143] Dunis, C. L., G. Giorgioni, J. Laws and J. Rudy, March 2010, Statistical Arbitrage and High-Frequency Data with an Application to Eurostoxx 50 Equities. Working Paper, Liverpool Business School. http://www.ljmu.ac.uk/ImagesEveryone/ Jozeflst(1).pdf [0144] Engle, R. and C. Granger, 1987, Co-integration and Error Correction: Representation, Estimation, and Testing. Econometrica 55, (2), pp. 251-276 [0145] Luenberger, D. G., A Correlation Pricing Formula, Journal of Economic Dynamics and Control, July 26, 2002, 1113-1126. [0146] Nath, P., 2003, High Frequency Pairs Trading with U.S Treasury Securities: Risks and Rewards for Hedge Funds, working paper (London Business School). [0147] Schizas, P., D. Thomakos and T. Wang, July 2011, Pairs Trading on International ETFs, Conference on Research on Economic Theory and Econometrics. 34

Claims

1. A smart statistical arbitrage system with dynamic trading strategies that change automatically according to the real-time market conditions, said system comprising: a real-time data collector, which receives real-time data from an external source and continuously transmits the same data to a database and other analytical components for real-time calculations of dynamic trading parameters; a database that stores a plurality of predetermined trading pairs and all data of financial instruments, including all related and processed historical data and real-time data; and an analytical module that includes a number of analytical components, wherein all kinds of dynamic trading parameters are computed and stored in a memory or in said database at a frequency of a predetermined time interval.

2. The smart system of claim 1, wherein all trading financial products and pairs in said database can be of any financial instruments as long as they are hedgeable; the selection and the combination of these financial instruments and pairs may be predetermined with or without knowledge of their historical characteristics.

3. The smart system of claim 1, wherein said database saves and stores all the related and processed data at a low frequency of a predetermined long time interval; said data of each trading instrument must include a maximum price, a minimum price and a closing price for each long time interval, wherein all said prices are the average of a bid price and an ask price.

4. The smart system of claim 1, wherein said low frequency of a predetermined long time interval can be longer than or equal to 24-hours, or any other time such as 8 hours, 4-hours, 2-hours or 1-hour; a high frequency of a predetermined short time interval used for parameter calculations can be longer than or equal to 60-seconds, or 35 WO 2013/071330 PCT/AU2011/001475 any other time such as 20-seconds, 10-seconds, 5-seconds or 3-seconds; a long-term refers to a fixed term of one year or any other predetermined periods; and a short term refers to a fixed term of one month or any other predetermined periods.

5. The smart system of claim 1, wherein a method for calculating the co-reliance rate, which is a quantifying measure of the degree of interdependency in a linear non statistical relationship of two random trading instruments, said method comprising the steps of: al) obtaining a max-change-rate for each instrument in a pair by deducting the close price of the previous time interval from the current maximum price, and then dividing the result by said close price of the previous time interval; a2) comparing the absolute value of max-change-rate of instrument 1 with the same of instrument 2 in said pair; a3) if the absolute value of max-change-rate of instrument 1 is larger than the same of instrument 2, then a co-reliance-max-rate equals to the ratio of max-change-rate of instrument 2 to the same of instrument 1; a4) if the absolute value of max-change-rate of instrument 2 is larger than the same of instrument 1, then said co-reliance-max-rate equals to the ratio of max-change-rate of instrument 1 to the same of instrument 2; b1) obtaining a min-change-rate for each instrument in said pair by deducting said close price of the previous time interval from the current minimum price, and then dividing the result by said close price of the previous time interval; b2) comparing the absolute value of min-change-rate of instrument 1 with the same of instrument 2 in said pair; 36 WO 2013/071330 PCT/AU2011/001475 b3) if the absolute value of min-change-rate of instrument 1 is larger than the same of instrument 2, then a co-reliance-min-rate equals to the ratio of min-change-rate of instrument 2 to the same of instrument 1; b4) if the absolute value of min-change-rate of instrument 2 is larger than the same of instrument 1, then said co-reliance-min-rate equals to the ratio of min-change-rate of instrument 1 to the same of instrument 2; c1) obtaining a close-change-rate for each instrument in said pair by deducting said close price of the previous time interval from the current close price, and then dividing the result by said close price of the previous time interval; c2) comparing the absolute value of close-change-rate of instrument 1 with the same of instrument 2 in said pair; c3) if the absolute value of close-change-rate of instrument 1 is larger than the same of instrument 2, then a co-reliance-close-rate equals to the ratio of close-change-rate of instrument 2 to the same of instrument 1; c4) if the absolute value of close-change-rate of instrument 2 is larger than the same of instrument 1, then said co-reliance-close-rate equals to the ratio of close-change rate of instrument 1 to the same of instrument 2; and d) obtaining said co-reliance rate of the current time interval for a pair through adding co-reliance-max-rate with co-reliance-min-rate and co-reliance-close-rate, and then dividing the result by 3.

6. The smart system of claim 1, wherein an analytical component within said analytical module utilises the method of claim 5 to calculate said co-reliance rate for each trading pair at said low frequency of a long time interval, then stores said calculated co-reliance rate as part of the dynamic parameters and historical data. 37 WO 2013/071330 PCT/AU2011/001475

7. The smart system of claim 6, wherein for each trading pair, an analytical component within said analytical module applies said saved co-reliance rate to calculate a long term moving average of co-reliance rate and a short-term moving average of co reliance rate, and then saves these calculated results as part of the dynamic parameters for making trading strategies.

8. The smart system of claim 1, wherein an analytical component within said analytical module calculates the price change rate of each trading instrument by using the maximum price, the minimum price and the closing price at said low frequency of a long time interval, then saves said calculated price change rate as part of the dynamic parameters and historical data.

9. The smart system of claim 8, wherein for each trading instrument, an analytical component within said analytical module applies said saved price change rate to calculate a long-term moving average of price change rate and a short-term moving average of price change rate, and then saves the calculated results as part of the dynamic parameters for making trading strategies.

10. The smart system of claim 1, wherein an analytical component within said analytical module calculates the long-term price range factor of each trading instrument by using the maximum price and the minimum price at said high frequency of a short time interval, and then saves the calculated price range factor as part of the dynamic parameters for making trading strategies.

11. The smart system of claim 1, wherein for each trading instrument, an analytical component within said analytical module calculates the short-term ratio of the current price deviation to the standard deviation by using the real-time price at said high frequency of a short time interval, and then saves the calculated ratio value as part of the dynamic parameters for making trading strategies. 38 WO 2013/071330 PCT/AU2011/001475

12. The system of claim 1, wherein a method for creating smart and dynamic trading strategies comprising the steps of: a) establishing an empirical calculation formula for said anti-trade strength, which includes various dynamic parameters that affect the profitability in trading; b) establishing an empirical calculation formula for said pro-trade strength, which includes dynamic real-time parameters that reflect the strength of the mean reversion; and c) establishing a dynamic trading prerequisite - said pro-trade strength is larger than said anti-trade strength.

13. The smart system of claim 1, wherein a method for calculating the pro-trade strength comprising the steps of: a) deducting said ratio of the current deviation to the standard deviation of the instrument 2 from the same of the instrument 1 in a pair; b) placing an exponential factor tP to the absolute value of the result calculated from above step a); c) multiplying a positive constant r to the result gained from above step b); d) deducting a positive constant A from the result gained from above step c) to obtain the final result, which will be the pro-trade strength for the time of the calculation; e) repeating the above steps a), b), c) and d) at said high frequency of a short time interval; and f) updating and saving the final result at each calculation as part of the dynamic parameters for making trading strategies. 39 WO 2013/071330 PCT/AU2011/001475

14. The smart system of claim 1, wherein a method for computing the current overall market trend strength, which measures a quantifying degree of the market trending, comprising the steps of: a) adding all of said ratios of the current deviation to the standard deviation of all trading instruments; b) dividing the above result by the number of the total trading instruments to obtain the outcome which will be the current market trend strength for the time of the calculation; c) repeating the above steps a) and b) at said high frequency of a short time interval; and d) updating and saving the final result at each calculation as part of the dynamic parameters for making trading strategies.

15. The smart system of claim 1, wherein a method for computing the anti-trade strength, comprising the steps of: a) calculating the weighting generated from both said long-term moving average and said short-term moving average of said price change rate of instrument 1 in a pair by applying empirical constant factors to determine the effects of said moving averages; b) calculating the weighting generated from both said long-term moving average and said short-term moving average of said co-reliance rate of said pair by applying empirical constant factors to determine the effects of said moving averages; c) calculating the weighting generated from both said long-term price range factors of instrument 1 and instrument 2 in said pair by applying empirical constants to determine the effects of said price range factors; 40 WO 2013/071330 PCT/AU2011/001475 d) calculating the weighting generated from both said long-term moving average and said short-term moving average of said price change rate of instrument 2 in said pair by applying empirical constant factors to determine the effects of said moving averages; e) calculating the weighting generated from said current market trend strength by applying empirical constant factors to determine the effect of said trend strength; f) adding all weightings of above steps of a), b), c), d) and e) to achieve a final result of the anti-trade strength. g) repeating all above steps at said high frequency of a short time interval; and h) updating and saving the final result at each calculation as part of the dynamic parameters for making trading strategies.

16. The smart system of claim 15, wherein a method for calculating said weighting generated from both said long-term and said short-term moving averages of said price change rate of an instrument in a pair, comprising the steps of: a) deducting the value of a positive constant a from said long-term moving average of said price change rate of said instrument in said pair; b) multiplying 100 by the absolute value of the result of above step a); c) placing an exponential factor 0 to the result calculated from above step b); d) multiplying the value of a positive constant 6 by the result of above step a); e) multiplying the result of above step c) by the result of above step d); 41 WO 2013/071330 PCT/AU2011/001475 f) deducting the value of a positive constant y from said short-term moving average of said price change rate of said instrument in said pair; g) multiplying 100 by the absolute value of the result of above step f); h) placing an exponential factor A to the result calculated from above step g); i) multiplying the value of a positive constant ( by the result of above step f); j) multiplying the result of above step h) by the result of above step i); and k) adding the result of above step e) with the result of above step j) to obtain the final result.

17. The smart system of claim 15, wherein a method for calculating said weighting generated from both said long-term and said short-term moving averages of said co reliance rate of a pair, comprising the steps of: a) multiplying 100 by said long-term moving average of said co-reliance rate of said pair; b) deducting the result of above step a) from a positive constant (; c) placing an exponential factor p to the result calculated from above step b); d) multiplying the value of a positive constant b by the result of above step b); e) multiplying the result of above step c) by the result of above step d); f) multiplying 100 by said short-term moving average of said co-reliance rate of said pair; g) deducting the result of above step f) from a positive constant p; 42 WO 2013/071330 PCT/AU2011/001475 h) placing an exponential factor q to the result calculated from above step g); i) multiplying the value of a positive constant A by the result of above step g); j) multiplying the result of above step h) by the result of above step i); and k) adding the result of above step e) with the result of above step j) to obtain the final result.

18. The smart system of claim 15, wherein a method for calculating said weighting generated from said price range factor of both instruments in a pair, comprising the steps of: a) deducting the value of a positive constant - from said price range factor of instrument 1 in said pair; b) placing an exponential factor E to the absolute value of the result calculated from above step a); c) multiplying the result of above step a) by the result of above step b); d) deducting the value of a positive constant T from said price range factor of instrument 2 in said pair; e) placing an exponential factor E to the absolute value of the result calculated from above step d); f) multiplying the result of above step d) by the result of above step e); and g) adding the result of above step c) with the result of above step f) to obtain the final result. 43 WO 2013/071330 PCT/AU2011/001475

19. The smart system of claim 15, wherein a method for calculating said weighting generated from said current market trend strength, comprising the steps of: a) deducting the value of a positive constant a from the absolute value of said current market trend strength; b) placing an exponential factor E to the absolute value of the result calculated from above step a); c) multiplying the value of a positive constant ) by the result of above step a); and d) multiplying the result of above step b) by the result of above step c) to obtain the final result.

20. The smart system of claim 1, wherein an analytical component within said analytical module compares said pro-trade strength with said anti-trade strength in order to make a trading decision at said high frequency of a short time interval; said system triggers a decision to open trades if said pro-trade strength is larger than said anti-trade strength, otherwise no action will be taken.

21. The smart system of claim 20, wherein if said system triggers a decision to open trades, said analytical module issues trading orders to sell instrument 1 and buy instrument 2 if instrument 1 has a larger ratio of the current price deviation to the standard deviation than the same of instrument 2 in said trading pair; or said analytical module issues trading orders to buy instrument 1 and sell instrument 2 if instrument 2 has a larger said ratio than the same of instrument 1 in said pair.

22. The smart system of claim I further comprising an automatic trading module that immediately executes trading orders received from the analytical module on a real time basis, wherein said trading orders include a buying order and a selling order. 44 WO 2013/071330 PCT/AU2011/001475 calculation of said pro-trade strength has a variation under one condition: when said ratio of the current price deviation to the standard deviation of said hedging instrument is smaller than the same of said hedged instrument, then a constant factor ri used for the calculation of said pro-trade strength will be altered from a positive value to a negative value; said hedging instrument is then concluded as not suitable to protect said hedged instrument at the time of calculation.

27. The hedging system of claim 24, wherein at said high frequency of a short time interval, an analytical component within said analytical module calculates the pro hedge strength by deducting the value of said anti-trade strength from the value of said pro-trade strength; if the result of said pro-hedge strength is larger than a preset value, which is set as an optimal point to execute the hedging, then the system will trigger an alert and display the time and the name of said hedging instrument and said hedged instrument with the value of said pro-hedge strength.

28. The hedging system of claim 24, wherein at any time a user can request a hedging suitability ranking list by typing the name of said hedged instrument and clicking on a submit button; and said analytical module will immediately search all said hedging instruments that pair with said hedged instrument, calculate all said pro-hedge strength values based on all real-time prices and other dynamic parameters and current market conditions, and finally list all the names of hedging instruments with a ranking sequence from the highest value of pro-hedge strength to the lowest.

29. The hedging system of claim 28, wherein a method for protecting an existing hedgeable investment that acts as said hedged instrument in said hedging system during a downtrend period, said method comprising the steps of: a) running said hedging system of claim 25 to obtain said hedging suitability ranking list; 46 WO 2013/071330 PCT/AU2011/001475 b) selling at least one hedging instrument from the top of said hedging suitability ranking list; c) re-running said hedging system of claim 25 to update said hedging suitability ranking list; d) determining if the market continues in the downtrend direction by evaluating said current market trend strength; e) buying back said hedging instruments that was sold previously, closing all hedging positions and taking profits; f) re-selling other hedging instruments selected from the top of said current hedging suitability ranking list; g) repeating above steps c), d), e) and f) again and again until the market changes to an uptrend direction; and h) repeating above steps c), d) and e) without f) when the market changes to an uptrend direction. 47 WO 2013/071330 PCT/AU2011/001475 AMENDED CLAIMS received by the International Bureau on 13 April 2012 (13.04.2012) 1. A smart statistical arbitrage system with dynamic trading strategies that change automatically according to the real-time market conditions, said system comprising: a real-time data collector, which receives real-time data from an external source and continuously transmits the same data to a database and other analytical components for real-time calculations of dynamic trading parameters; a database that stores a plurality of predetermined trading pairs and data of financial instruments, said data specifically including a co-reliance rate, a pro-trade strength, a market trend strength, an anti-trade strength and other related and processed historical data and real-time data; and an analytical module that includes a number of analytical components, wherein dynamic trading parameters are computed and stored in a memory or in said database at a frequency of a predetermined time interval, said dynamic trading parameters particularly include said co-reliance rate with its moving averages, said pro-trade strength, said market trend strength and said anti-trade strength. 2. The smart system of claim 1, wherein a method for calculating the co-reliance rate, which is a quantifying measure of the degree of interdependency in a linear non statistical relationship of two random trading instruments, said method comprising the steps of: al) obtaining a max-change-rate for each instrument in a pair by deducting the close price of the previous time interval from the current maximum price, and then dividing the result by said close price of the previous time interval; a2) comparing the absolute value of max-change-rate of instrument 1 with the same of instrument 2 in said pair; 48 WO 2013/071330 PCT/AU2011/001475 a3) if the absolute value of max-change-rate of instrument 1 is larger than the same of instrument 2, then a co-reliance-max-rate equals to the ratio of max-change-rate of instrument 2 to the same of instrument 1; a4) if the absolute value of max-change-rate of instrument 2 is larger than the same of instrument 1, then said co-reliance-max-rate equals to the ratio of max-change-rate of instrument 1 to the same of instrument 2; b1) obtaining a min-change-rate for each instrument in said pair by deducting said close price of the previous time interval from the current minimum price, and then dividing the result by said close price of the previous time interval; b2) comparing the absolute value of min-change-rate of instrument 1 with the same of instrument 2 in said pair; b3) if the absolute value of min-change-rate of instrument 1 is larger than the same of instrument 2, then a co-reliance-min-rate equals to the ratio of min-change-rate of instrument 2 to the same of instrument 1; b4) if the absolute value of min-change-rate of instrument 2 is larger than the same of instrument 1, then said co-reliance-min-rate equals to the ratio of min-change-rate of instrument 1 to the same of instrument 2; c1) obtaining a close-change-rate for each instrument in said pair by deducting said close price of the previous time interval from the current close price, and then dividing the result by said close price of the previous time interval; c2) comparing the absolute value of close-change-rate of instrument 1 with the same of instrument 2 in said pair; 49 WO 2013/071330 PCT/AU2011/001475 c3) if the absolute value of close-change-rate of instrument 1is larger than the same of instrument 2, then a co-reliance-close-rate equals to the ratio of close-change-rate of instrument 2 to the same of instrument 1; c4) if the absolute value of close-change-rate of instrument 2 is larger than the same of instrument 1, then said co-reliance-close-rate equals to the ratio of close-change rate of instrument 1 to the same of instrument 2; and d) obtaining said co-reliance rate of the current time interval for a pair through adding co-reliance-max-rate with co-reliance-min-rate and co-reliance-close-rate, and then dividing the result by 3. 3. The smart system of claim 1, wherein an analytical component within said analytical module utilises the method of claim 2 to calculate said co-reliance rate for each trading pair at a low frequency of a long time interval, then stores said calculated co reliance rate as part of the dynamic parameters and historical data. 4. The smart system of claim 3, wherein for each trading pair, an analytical component within said analytical module applies said saved co-reliance rate to calculate a long term moving average of co-reliance rate and a short-term moving average of co reliance rate, and then saves these calculated results as part of the dynamic parameters for making trading strategies. 5. The system of claim 1, wherein a method for creating smart and dynamic trading strategies comprising the steps of: a) establishing an empirical calculation formula for said anti-trade strength, which includes various dynamic parameters that affect the profitability in trading; 50 WO 2013/071330 PCT/AU2011/001475 b) establishing an empirical calculation formula for said pro-trade strength, which includes dynamic real-time parameters that reflect the strength of the mean reversion; and c) establishing a dynamic trading prerequisite - said pro-trade strength is larger than said anti-trade strength. 6. The smart system of claim 1, wherein a method for calculating the pro-trade strength comprising the steps of: a) deducting a ratio of the current deviation to the standard deviation of the instrument 2 from the same of the instrument 1 in a pair; b) placing an exponential factor $ to the absolute value of the result calculated from above step a); c) multiplying a positive constant r to the result gained from above step b); d) deducting a positive constant A from the result gained from above step c) to obtain the final result, which will be the pro-trade strength for the time of the calculation; e) repeating the above steps a), b), c) and d) at a high frequency of a short time interval; and f) updating and saving the final result at each calculation as part of the dynamic parameters for making trading strategies. 7. The smart system of claim 1, wherein a method for computing the market trend strength, which measures a quantifying degree of the market trending, comprising the steps of: 51 WO 2013/071330 PCT/AU2011/001475 a) adding all of said ratios of the current deviation to the standard deviation of all trading instruments; b) dividing the above result by the number of the total trading instruments to obtain the outcome which will be the market trend strength for the time of the calculation; c) repeating the above steps a) and b) at a high frequency of a short time interval; and d) updating and saving the final result at each calculation as part of the dynamic parameters for making trading strategies. 8. The smart system of claim 1, wherein a method for computing the anti-trade strength, comprising the steps of: a) calculating the weighting generated from both a long-term moving average and a short-term moving average of a price change rate of instrument 1 in a pair by applying empirical constant factors to determine the effects of said moving averages; b) calculating the weighting generated from both said long-term moving average and said short-term moving average of said co-reliance rate of said pair by applying empirical constant factors to determine the effects of said moving averages; c) calculating the weighting generated from both long-term price range factors of instrument 1 and instrument 2 in said pair by applying empirical constants to determine the effects of said price range factors; d) calculating the weighting generated from both a long-term moving average and a short-term moving average of a price change rate of instrument 2 in said pair by applying empirical constant factors to determine the effects of said moving averages; e) calculating the weighting generated from said market trend strength by applying empirical constant factors to determine the effect of said trend strength; 52 WO 2013/071330 PCT/AU2011/001475 f) adding all weightings of above steps of a), b), c), d) and e) to achieve a final result of the anti-trade strength. g) repeating all above steps at a high frequency of a short time interval; and h) updating and saving the final result at each calculation as part of the dynamic parameters for making trading strategies. 9. The smart system of claim 8, wherein a method for calculating said weighting generated from both said long-term and said short-term moving averages of said price change rate of an instrument in a pair, comprising the steps of: a) deducting the value of a positive constant a from said long-term moving average of said price change rate of said instrument in said pair; b) multiplying 100 by the absolute value of the result of above step a); c) placing an exponential factor 0 to the result calculated from above step b); d) multiplying the value of a positive constant 6 by the result of above step a); e) multiplying the result of above step c) by the result of above step d); f) deducting the value of a positive constant y from said short-term moving average of said price change rate of said instrument in said pair; g) multiplying 100 by the absolute value of the result of above step f); h) placing an exponential factor X to the result calculated from above step g); i) multiplying the value of a positive constant ( by the result of above step f); j) multiplying the result of above step h) by the result of above step i); and 53 WO 2013/071330 PCT/AU2011/001475 k) adding the result of above step e) with the result of above step j) to obtain the final result. 10. The smart system of claim 8, wherein a method for calculating said weighting generated from both said long-term and said short-term moving averages of said co reliance rate of a pair, comprising the steps of: a) multiplying 100 by said long-term moving average of said co-reliance rate of said pair; b) deducting the result of above step a) from a positive constant (; c) placing an exponential factor t to the result calculated from above step b); d) multiplying the value of a positive constant 4) by the result of above step b); e) multiplying the result of above step c) by the result of above step d); f) multiplying 100 by said short-term moving average of said co-reliance rate of said pair; g) deducting the result of above step f) from a positive constant p; h) placing an exponential factor q to the result calculated from above step g); i) multiplying the value of a positive constant A by the result of above step g); j) multiplying the result of above step h) by the result of above step i); and k) adding the result of above step e) with the result of above step j) to obtain the final result. 54 WO 2013/071330 PCT/AU2011/001475 11. The smart system of claim 8, wherein a method for calculating said weighting generated from said price range factors of both instruments in a pair, comprising the steps of: a) deducting the value of a positive constant t from said price range factor of instrument 1 in said pair; b) placing an exponential factor E to the absolute value of the result calculated from above step a); c) multiplying the result of above step a) by the result of above step b); d) deducting the value of a positive constant T from said price range factor of instrument 2 in said pair; e) placing an exponential factor E to the absolute value of the result calculated from above step d); f) multiplying the result of above step d) by the result of above step e); and g) adding the result of above step c) with the result of above step f) to obtain the final result. 12. The smart system of claim 8, wherein a method for calculating said weighting generated from said market trend strength, comprising the steps of: a) deducting the value of a positive constant a from the absolute value of said market trend strength; b) placing an exponential factor E to the absolute value of the result calculated from above step a); 55 A RA KIM ME ^I 1 ~ I A MFI/I A4 ^I WO 2013/071330 PCT/AU2011/001475 c) multiplying the value of a positive constant 0 by the result of above step a); and d) multiplying the result of above step b) by the result of above step c) to obtain the final result. 13. The smart system of claim 1, wherein an analytical component within said analytical module compares said pro-trade strength with said anti-trade strength in order to make a trading decision at a high frequency of a short time interval; said system triggers a decision to open trades if said pro-trade strength is larger than said anti-trade strength, otherwise no action will be taken. 14. The smart system of claim 13, wherein if said system triggers a decision to open trades, said analytical module issues trading orders to sell instrument 1 and buy instrument 2 if instrument 1 has a larger ratio of the current price deviation to the standard deviation than the same of instrument 2 in said trading pair; or said analytical module issues trading orders to buy instrument 1 and sell instrument 2 if instrument 2 has a larger said ratio than the same of instrument 1 in said pair. 15. A smart hedging system with dynamic hedging strategies that change automatically according to the real-time market conditions, said hedging system comprising: a real-time data collector, which receives real-time data from an external source and continuously transmits the same data to a database and other analytical components for real-time calculations of dynamic hedging parameters; a database that stores a plurality of hedging pairs and data of financial instruments, said data specifically including a co-reliance rate, a pro-trade strength, a market trend strength, an anti-trade strength, a pro-hedge strength and other related and processed historical data and real-time data; 56 WO 2013/071330 PCT/AU2011/001475 an analytical module that includes a number of analytical components, wherein dynamic hedging parameters are computed and stored in a memory or in the database at a frequency of a predetermined time interval, said dynamic hedging parameters particularly include said co-reliance rate with its moving averages, said pro-trade strength, said market trend strength, said anti-trade strength and said pro-hedge strength; and a displaying module, which shows a ranking list with values of pro-hedge strength from the highest to the lowest received from the analytical module, when the ranking list is requested. 16. The hedging system of claim 15, wherein the financial instruments that have already been purchased as existing investments and require protection are the hedged instruments, which are stored in the database together with the hedging instruments that are preselected for the purpose to hedge against the hedged instruments in an attempt to protect the existing investments; a pre-combined pair of instruments in said database must consist of one hedging instrument and one hedged instrument. 17. The hedging system of claim 15, wherein said hedging system applies all parameters that are identical to said smart system of claim 1, except that the calculation of said pro-trade strength has a variation under one condition: when a ratio of the current price deviation to the standard deviation of said hedging instrument is smaller than the same of said hedged instrument, then a constant factor n used for the calculation of said pro-trade strength will be altered from a positive value to a negative value; said hedging instrument is then concluded as not suitable to protect said hedged instrument at the time of calculation. 18. The hedging system of claim 15, wherein at a high frequency of a short time interval, an analytical component within said analytical module calculates the pro hedge strength by deducting the value of said anti-trade strength from the value of 57 WO 2013/071330 PCT/AU2011/001475 said pro-trade strength; if the result of said pro-hedge strength is larger than a preset value, which is set as an optimal point to execute the hedging, then the system will trigger an alert and display the time and the name of said hedging instrument and said hedged instrument with the value of said pro-hedge strength. 19. The hedging system of claim 15, wherein at any time a user can request a hedging suitability ranking list by typing the name of said hedged instrument and clicking on a submit button; and said analytical module will immediately search all said hedging instruments that pair with said hedged instrument, calculate all said pro-hedge strength values based on all real-time prices and other dynamic parameters and current market conditions, and finally list all the names of hedging instruments with a ranking sequence from the highest value of pro-hedge strength to the lowest. 20. The hedging system of claim 19, wherein a method for protecting an existing hedgeable investment that acts as said hedged instrument in said hedging system during a downtrend period, said method comprising the steps of: a) running said hedging system of claim 25 to obtain said hedging suitability ranking list; b) selling at least one hedging instrument from the top of said hedging suitability ranking list; c) re-running said hedging system of claim 25 to update said hedging suitability ranking list; d) determining if the market continues in the downtrend direction by evaluating said market trend strength; e) buying back said hedging instruments that was sold previously, closing all hedging positions and taking profits; 58 AMENDED SHEET (ARTICLE 19) WO 2013/071330 PCT/AU2011/001475 f) re-selling other hedging instruments selected from the top of said current hedging suitability ranking list; g) repeating above steps c), d), e) and f) again and again until the market changes to an uptrend direction; and h) repeating above steps c), d) and e) without f) when the market changes to an uptrend direction. 59