Let’s say you’ve got a long only trend following system. A long only trend following system is not going to do particularly well in a sideways market or a downwards market. That’s a fact of life. It is designed specifically to profit extensively when the market is trending up. Again, this comes back to why your system works, and it won’t work if the market is going sideways or the market is going down –but that does not mean it’s broken. It’s like buying a Porsche, driving it out to the beach, and then taking it back to the dealership and saying, “Mate, this Porsche doesn’t work on the beach!”, I mean you just wouldn’t do that. It’s not designed to go on the beach, and same with a long only trend following system. It is designed to make money when the market is trending higher –which most stock markets tend to trend over the longer term to the up side.
There are certainly periods of time where it won’t and that does not mean the system is broken. It means you’ll go into drawdown. So when your system is out of sync at the market, you go into drawdown. That’s the payoff.
If you can’t ask or answer the question, “Why is my system not making money?”, then you have a problem. So if I’m trading a trend following system and I’m in drawdown, but I can see that the market is moving sideways, then I have nothing to worry about, that’s just a fact of life. But if the market is trending up and I’m using a trend following system, and I’m losing money, well then you’ve got a serious question right there. And if you can’t answer that question, well then you’ve got a problem, and you probably should stop trading it.
Technically what we do is, we’re constantly running simulations side by side with the real time trading – I literally do it weekly. I want to see that I’m getting the same number of trades, the same kind of general return that I’m getting every single week from the simulation and from my real time trading.
I’ll give you an example of where this has helped me in the past. It was about 2011, I was trading a shorter term momentum style strategy and my real time results were diverging from my simulated results. I did all the calculations, I went all the way back and had a look at about three years worth of trading. I realized that slippage had started to increase, so much so that it was costing me 11% per annum, and that 11% basically meant that the system was not viable anymore. That was a liquidity issue rather than a system issue. I worked hard to try and find a solution but I couldn’t. Basically, one of the things we’ve seen in the Australian market since the GFC is a significant drop in liquidity. That increased slippage so much so, it invalidated my strategy. So that’s a real time example of where the strategy was no longer tradable. I could have taken that strategy and put it into the U.S. market where there’s a lot more liquidity but I had the US market covered.
So we’re always running simulations against the real time results. If there is a divergence, you’ve got to try and figure out why. If you can figure out why then you’re in a better position. If it was a one off event such 1997 stock market crash, well there’s probably nothing intrinsically wrong with the strategy. It’s probably just one of those one off events that will impact everybody. What you don’t want to see is a slow divergence away from the simulated results over time. So you will have set benchmarks using your historical performance metrics and when you start to deviate away from those then you’ve probably got a problem.