Building Mechanical Trading System - Teguh Pranoto Chen

1st-Nov-12 TASS at SCCCI  6.30pm - 9.30pm   (about 60 attended)

About Mechanical Trading Systems
Mechanical Trading Systems are precise sets of trading rules and designed to be mechanical in nature. It is to provide trading signals when prescribed conditions are fulfilled, e.g. when and at what price to enter or exit a trade. These systems are designed to be executed without taking consideration of human emotion.

About the Speaker:
Teguh Pranoto Chen specializes in Mechanical Trading Systems (MTS). Using advanced programming and statistical analysis, MTS is a powerful tool to profitability. Rarely known to average trader, mechanical trading is used by most market professionals. The presentation will share an introduction of Mechanical Trading System and how to build a Portfolio of Trading Systems.

Building Portfolio Trading System

Objectives:

Average Trader:  Maximise Profit
Season Trader:  Minimise Risk, Preserve Capital, Consistent Return with small Profit, once awhile make outsized return.

"There are bold traders, there are old traders, but there are no bold and old traders."

Mechanical Trading System

Precise sets of trading rules and designed to be mechanical in nature
To provide trading signals when prescribed conditions are fulfilled when at what price to enter or exit

Why Mechanical System?
Way to minimise risk by testing theory before risking capital
Diversify portfolio to smoothen equity progression

Disadvantages
Past performance is not an indication of the future
Take time and effort to learn the ropes
Prone to mistakes while converting good ideas into reality

Key Watch Out
Curve fitting - fit data with the system and see which one works well
Future leak - looking into future data to come up with past signals
Snake oil - black box system with no logic or code disclosed
Tradability and slippage

"If it is too good to be true, it if NOT TRUE"

3M of success
Methodology
Money Management
Mental - manage emotion while trading own system

Holy Grail
Accuracy
Frequency
Leverage

Building a Trading System
Market Condition:  High and Low Volatality
High Volatility - momentum trading combine with swing trading (momentum, oscillator and short period indicators)
Low Volatility - Trending following (MA cross over, channel and bands), low profit potential, options trading etc.

Time Frame:  Relative to different markets
Intra day - ticks to hours
Short term  few days to weeks
Medium term - few weeks to months
Long term - months to years

Instruments: Stock, ETF, Options, Futures, Commodities, Forex

Approaches:
Donchian breakout
Moving average crossover
Volatility open range breakout
Day trading
Pattern recognition

Automation
Manual order entry at starting point for system
Broker assistant required at certain conditions
Fully automation suitable for HFT (High Frequency Trading)

Optimization
Walk forward using past data to develop system and run sysem into future.
Nonoptimized: one set of variable throughout system life
Artificial intelligence

Key statistics:
K-ratio - smothenss of equity curve
Win-loss ratio
Average bars held
Draw down
Sharpe ratio - risk adjusted measure

Back test
Period analysis
Trade by trade analysis
Monte Carlo simulation
Risk of ruin

Tools of the trade
Amibroker
Trade station
Multicharts
Neuroshell
Excel
@Risk
Datasource EOD or RT

Backteset on Donchian Breakout (Turtle system)
Buy at breakout of 20-day high or Sell at break down of 10-day low.
Annual return 16% over 4 years, Draw down -20% but k-ratio and sharpe ration < 1%

Fixed lot size vs leverage upside, above test based on 1 standard fix lot size.

Size of Trading System
Starter  10k-30k  expect 100% pa but limited to max 3 instrument and possible to blow out ac
Mediium 30k - 50k
Full  50k - 100k
Global  >  100k  expect 30-40% pa

Portfolio Trading System
Comprises combination of systems
Covering different time frame
Spread over differnt instruments
Utilising different approaches