Trading System a new approach

Part I: A Practical Guide to Trading System

Development and Evaluation 1

Chapter 1: What is a trading system? 3

1.1 An easy example of a trading system 4

1.2 Why you need a trading system 5

1.3 The science of trading systems 7

Chapter 2: Design, test, optimisation and evaluation of a

trading system 9

2.1 Design 9

Getting started 9

The programming task 10

Which timeframe to trade? 11

2.2 Test 12

The importance of the market data 12

The length of your back-testing period 14

Rule complexity and degrees of freedom 15

2.3 The forecasting power of a trading system 19

Optimisation 19

Walk forward analysis 21

Robustness 23

2.4 Evaluation of a trading system 27

What to look for in an indicator 27

Average trade 28

Percentage of profitable trades 28

Profit factor 30

Drawdown 30

Time averages 31

RINA Index 32

2.5 Conclusion 33

Part II: Trading System Development and Evaluation

of a Real Case 35

Chapter 3: How to develop a trading system step-by-step –

using the example of the British pound/US dollar pair 37

Introduction 37

3.1 The birth of a trading system 38

The free LUXOR system code 39

The entry logic 41

3.2 First evaluation of the trading system 43

Calculation without slippage and commissions 43

Calculation after adding slippage and commissions 47

3.3 Variation of the input parameters: optimisation and

stability diagrams 49

What does stability of a system’s input parameter mean?

A short theoretical excursion 49

Dependency of main system figures on the two moving

averages 51

Result with optimised input values 56

3.4 Inserting an intraday time filter 59

Finding the best entry time 59

Result with added time filter 61

3.5 Determination of appropriate exits – risk management 64

The concept of Maximum Adverse Excursion (MAE) 66

Inserting a risk stop loss 70

Adding a trailing stop 74

Looking for profit targets: Maximum Favorable

Excursion (MFE) 76

Summary: Result of the entry logic with the three

added exits 79

How exits are affected by money management 83

3.6 Summary: Step-by-step development of a trading system 86

Chapter 4: Two methods for evaluating the system’s

predictive power 89

4.1 Timescale analysis 90

Changing the compression of the price data 90

LUXOR tested on different bar compressions 92

Net profit and maximum drawdown dependent on the

traded bar length 96

Explanation for the time dependency of the system 97

4.2 Monte Carlo analysis 101

The principle of Monte Carlo analysis 101

Exchanging the order of the performed trades 104

Probabilities and confidence levels 105

Performing a Monte Carlo analysis with the LUXOR

trading system 107

Limitations of the Monte Carlo method 108

Chapter 5: The factors around your system 111

5.1 The market’s long/short bias 112

The trend is your friend? 112

Consequences for system development 114

5.2 Out-of-sample deterioration 115

A Bollinger Band system with logic and code 115

Optimising the Bollinger Band system 118

Out-of-sample result 119

Reasons for the out-of-sample deterioration 121

5.3 The market data bias 122

Expanding the training period 122

Conclusion: How to choose your training data 126

5.4 Optimisation and over-fitting 126

Step-by-step optimisation of the LUXOR system 126

Results depending on the number of optimised parameters 127

The meaning of the trading system’s complexity 134

5.5 Rule complexity explained with polynomial curve fitting 136

Interpolating data points with polynomial functions 136

Predictive power of the different polynomials 142

Conclusions for trading system development 145

Chapter 6: Periodic re-optimisation and walk forward analysis 147

6.1 Short repetition: “normal”, static optimisation 147

6.2 Anchored vs. rolling walk forward analysis (WFA) 149

6.3 Rolling WFA on the LUXOR system 150

Periodic optimisation of the two main system parameters 150

Out-of-sample test result 153

Conclusion 155

6.4 The meaning of sample size and market structure 155

Chapter 7: Position sizing example, using the LUXOR system 159

7.1 Definitions: money management vs. risk management 159

Risk management (RM) 159

Money management (MM) 160

7.2 Application of different MM schemes 161

Reference: The system traded with one lot 162

Maximum drawdown MM 163

Fixed fractional MM 164

Fixed ratio MM 168

7.3 Monte Carlo analysis of the position sized system 173

7.4 Conclusion 175

Part III: Systematic Portfolio Trading 177

Chapter 8: Dynamic portfolio construction 179

8.1 Introduction to portfolio construction 179

A list of the main available software 180

The role of correlations 181

Publications and theoretical tools 182

Portfolio trading in practice 183

Total vs. partial equity contribution 185

8.2 Correlation among equity lines 186

8.3 A dynamic approach: equity line crossover 188

8.4 Dynamic portfolio composition: the walk forward analysis activator 190

8.5 Largest losing trade/largest losing streak/largest drawdown 192

Conclusion 193

Appendices: Systems and ideas 199

Appendix 1: Bollinger Band system 201

1.1 Idea 201

1.2 Entry logic and Easy Language code 202

1.3 Application of the strategy to seven markets with

same parameters 204

1.4 Results and conclusions 204

Appendix 2: The Triangle system 209

2.1 Idea 209

2.2 Programming and coding 210

2.3 Application to different liquid futures markets with

same parameters 211

2.4 Advantages in building a portfolio 212

2.5 Conclusion