I keep track of my trading system performance with R multiples. I examine expectancy and accuracy of my system. This is a common practice applied by many other traders as well.
Something is missing here. When I trade ‘my system’ or ‘my plan’, it should be something that fits my risk appetite, my personality and my time of trading (morning, afternoon or like me, in the middle of the night) etc. So, I think it is neccessary to integrate these elements when evaluating a system.
How am I going to do that? Mechanical backtesting is just part of evaluation and it means a little. No emotion was involved. Some dumb numbers from backtesting sytem give me false sense of security and confidence. We can obtain some statistics from backtesting, that’s true. For example, you can get expected drawdown amount, number of continuous losses, expectancy, accuracy etc. The hard, cold figures.
Trading is an action by trader. Trader is the center of the system. And trader, I am referring to human (creatures that sometimes act like monkey, but will never admit they are relatives of monkey). Same system will have different performance when it is operated by different traders.
I like forward testing. I trade the real money, with small amount of money, for example, in futures trading, I would only trade 1 contract. I will try to carry this testing for 100 trades. Try my very best to follow the trading signals. During these 100 trades, I will meet problems like tiredness when trading at 1 am in the morning or unable to take losses etc.
If backtesting shows me that there will be a string of 10 losses, I might feel ‘No big deal’. But when it comes to real money, I might have chickened out after a string of 5 losses. So , in reality, the system does not suit me, even though I felt ‘No big deal’ when backtesting told me I shall expect a string of 10 losses.
So, here is the summary of how do I evaluate my trading system:
- Mechanical backtesting to obtain figures such as expectancy, accuracy, maximum draw down and number of continuus losses. (optional, since you will have a real result later)
- Trading 1 contract by strictly following the sytem signal for 100 trades. (The more the merrier.This is a chance to test your discipline)
- Recording down the details of real trades (amount earned, lost, and R-multiples, calculate the expectancy)
- Stop trading if maximum draw down has hit your comfort level, I mean financially. This is a strong signal that telling me the system might not suit me well.
- I will be very cautious if the system shows me negative expectancy after 70 trades. And I will stop trading and review the system if it shows me negative expectancy after 100 trades.
This might be nonsense to some people. But I like to focus on psychologcial aspect of trading. If we agree that trader is part of the system, then forward testing is crucial when evaluating a system. A trading system which can be operated well and with confidence by me is the best system for me.
powered by performancing firefox