Richard Dennis once said that even if he published his trading rules in the newspaper, people wouldn’t follow them. With that idea in mind, let’s take a look at a trading system that actually has been published and widely distributed.
About The System
One of the most distributed trading systems of all time can be found in the appendices of Michael Covel’s Trend Following. It is provided as a real world example of how a trend following system works. There is even some programming language included.
The system is designed to establish a position on an 89 day price breakout and exit that position on a 13 day price breakdown. It also uses a position sizing algorithm that factors in the dollar value that is to be risked on each trade but also makes sure that each trade risks less than 2% of the account. Backtesting was done across 15 commodity and currency markets over a ten year period.
The backtesting results show us that this simple system returned an average of better than 21% annually, while only profiting on 42.53% of its trades. The average winning trade was a little more than twice the average losing trade, and the maximum drawdown was 23.05%. The system has a Sharpe Ratio of 1.02, which is good, but not great.
Like most of the systems I have looked at in this series, this would be a good starting point if you were developing a system from scratch. The system clearly has an edge, but when you factor in commissions and slippage that edge will shrink. There is also the constant threat of a Black Swan event destroying your entire portfolio.
The strong points of this system are that it is built to trade multiple markets, has a position sizing algorithm, and that is can be profitable while winning on less than half of its trades. The biggest negative of this system is that we only have 10 years of backtesting data. It would be interesting to see how the system fared from 2001 through 2012.
Position Sizing Algorithm
The most interesting aspect of this system is its position sizing algorithm. It is designed to make sure that each position is less than 2% of the portfolio, but also adjusts those positions based on the dollar value of the risk they are exposed to. The system does this by calculating the position size in two separate ways and then using the more conservative result.
The first approach simply calculates 2% of the total equity and then divides that number by the dollar value of the entry minus the stop, which is the amount being risked. This approach is just using the entry and stop to calculate how big the position can be to account for 2% of the total equity if it is unsuccessful and the stop is triggered.
The second approach takes 2% of the total equity and divides it by the dollar value of twice the average true range. This approach takes volatility into consideration, thus adding an extra layer of risk protection.
Trading breakout strategies can often seem far more complex than they actually are. Breakout systems are actually incredibly simple, however as traders, we tend to overcomplicate things. A Breakout of Breakdown is simply a new high or low for a given time period.
For example, this system uses an 89 day price breakout as an entry point. This means that the system would go long when a market hits an 89 day high, meaning the market is at its highest level for the past 89 days. In the opposite situation, where a market hits an 89 day low, the system would establish a short position.
When a position is taken, the system sets a stop at the 13 day high or low depending on the direction. If the position is long, the stop would be placed at the 13 day low. This stop value is then recalculated by the trading software every day.
On this chart of the SPY, the red lines represent the 89 day highs and lows. The blue lines represent the 13 day highs and lows. These lines make it easier to visualize the breakouts that the system is looking for.
The system would have generated a buy signal when it broke through to a new 89 day high at the beginning of 2013. Having the benefit of hindsight, this was an ideal time to establish a long position. When the long position was established, a stop would have been set at the 13 day low, which is the solid blue line. That stop would have been triggered at the end of February.
There would have been another buy signal at the beginning of March. That position would have run until it was stopped out in April. Then another long position would have been taken in early May.
This system would have captured a significant portion of this markets uptrend over the first few months of 2013. This is similar to the results we have seen from other trend following systems that we have looked at. While each of these systems goes about it in a slightly different way, they are all attempting to capture significant portions of long term trends.