Mechanical versus Discretionary Systems

After finally importing 2 more years of intraday data, my Dooku trading system has failed to backtest well using 4 years of intraday data.  Where does this leave me?  In a word, LOST.

I need to investigate furthur the recommended methods to backtest a trading system.  Of course I realize that past results don’t guarantee future results but I need to understand it more.  1 question I need to ask myself is:  Will my system will be totally mechanical or a discretionary system?

Here are some good pros and cons of both:

Mechanical systems

Advantages 1. This kind of system can be automated and backtested efficiently. 2. It has very rigid rules. Either, there is a trade or there isn’t. 3. Mechanical traders are less susceptible to emotions than discretionary traders. Disadvantages 1. Most traders backtest Forex trading systems incorrectly. In order to produce accurate results you need tick data. 2. The Forex market is always changing. The Forex market (and all markets) has a random component. The market conditions may look similar, but they are never the same. 3. A system that worked successfully the past year doesn’t necessary mean it will work this year.

Discretionary systems

Advantages 1. Discretionary systems are easily adaptable to new market conditions. 2. Trading decisions are based on experience. Traders learn to see which trading signals have higher probability of success. Disadvantages 1. They cannot be backtested or automated, since there is always a thought decision to be made. 2. It takes time to develop the experience required to trade successfully and track trades in a discretionary way. At early stages this can be dangerous.

Popularity: 3%

Trading System signals

I still have two short GBP/USD and EUR/USD positions from last week that were opened due to signals that my Dooku trading system generated.  The market hasn’t been doing much since then yet I’m sitting on an unrealized profit of $500. 

Yesterday I stated that my trading system had a profit target of 225 pips and a stop loss of 75 pips.  Though these are the stops and limits I place on each trade, my trading system can generate reversal signals that prevent my profit target from ever being hit.   The most I can lose on any given trade is 75 pips.  Generally if the trade moves at least 50 pips in my direction, I’m usually guaranteed that I will profit from at least these 50 pips.  It’s when it moves above this when I have a chance of realizing the 225 pip profit target as long as my trading system doesn’t generate a reversal signal.  In backtesting, my maximum profit target was not hit as often as my maximum stop loss but forward testing should give me more insight on how this trading system will perform in the long run.

I’ve been busy trying to import 5 years worth of intraday data into Amibroker for backtesting.  I know I was talking up Amibroker like it was the greatest thing since sliced bread but 1 drawdown is the fact that they use flat files for historical data as opposed to an optimized database.  What this means is that as more and more data is added, the slower and slower it is to import.  I figured the import of 5 years of 1 minute data for 10 currency pairs will take about 2 days for a single cpu pc at 100% utilization.  Unfortunately I don’t have a dual-cpu or dual core system right now though I’m very tempted to buy one. 

Popularity: 3%