Strange Currencies: A Donchian Trend Following System

The Trading System,

I call this trading system 'Strange Currencies' because it's a fairly ordinary mechanical trend following system with a rather strange exit (which seems to work nevertheless). Trend following systems typically produce more losing trades than winning ones, but the winners tend to out-weight the losers in terms of size. One might expect a good trend following system to produce a winning trade to losing trade ratio of around 60/40 with the 40% of winning trades (on average) winning 2 to 3 times as much a losing trade typically loses.

Trend following systems aim to take advantage of the big moves by buying high and selling even higher; or selling low and buying back even lower (shorting). A well-known example of such a system might be the simple Donchian 55/20 breakout system. The Donchian 55/20 breakout system either goes long when the price breaks it's 55 day high and closes the long position when the price breaks it's 20 day low; or it goes short when the price falls below it's 55 day low and closes the short position when the price rises above it's 20 day high. Buying high and selling low seems contrary to all common sense (most people want to buy low and sell high), but these systems work by taking full advantage of the few really big unforeseen moves up and down by getting in fairly early and hanging on (usually for months) until the trend is well and truly over. However, most of the time, I'd guess around 60% of the time if I had to put a figure on it, the moves up after the 55 day high is broken won't be large enough to pay for the loses that are occurred waiting for the 20 day low needed to close the position, so these systems are very hard to stick to and follow when a string of loses is producing a large draw down and the big move that's supposed to pay for it all seems elusive.

A modification I have come up with is to go long when the price closes higher than it has closed in x number of days and close the long position if it doesn't close at a new high within y number of days; or go short when the price closes lower than it has closed in x number of days and close the short position if it doesn't close even lower within y number of days. Exiting the position if it doesn't make a new extreme high or low within y number for days seems to work well as the trade is exited before to many pips are given back producing a higher percentage of winning trades; and if the market spends more than y number of days consolidating before moving again this system will simply exit and enter the market again as soon as the consolidation period is over, usually without to big a loss and almost certainly not the size of loss one would incur waiting for a 20 day low. These rules also mean the system will usually stay out of the market when it's not trending, which is a huge plus for any trend following system.

Stop losses. Adding stop losses to any system seems to make the system's expectancy worse unless the stop loss is very large; for this system I'm using a stop loss of 10% of the market price.

Recap: Rules for the 'Strange Currencies' Technical Trading System,

• Go long if the price closes higher than it has closed in x number of days.
• Close the long position if the price fails to close even higher in y number of days.
• Use a stop loss of 10% of the closing price on the day you went long.

• Go short if the price closes lower than it has closed in x number of days.
• Close the short position if the price fails to close even lower in y number of days.
• Use a stop loss of 10% of the closing price on the day you went short.

• X = 120, Y = 12 (longest).
• X = 60, Y = 10 (medium).
• X = 30, Y = 8 (shortest).

The Results,

The most obvious thing the results show us is the importance of using a suitable trading time frame, the larger the time frame we're using the more significant the high or low is. On the largest time frame we tested, 120 days, every currency pair produced winning trades of 50% or greater with the winnings being substantially bigger than the losses. To see the full details of the results right click and save target as here. Please note the number of days is the number of trading days and not the number of calendar days.

EUR/USD

USD/JPY

GBP/USD

USDCHF

USD/CAD

Suggested Improvements,

They say a good trade rarely goes to far against you, that is, a winning trade won't normally go to far into the red. Whilst adding a tight stop loss usually makes any system worse there is certainly room for improvement with stop losses here. The full results contain the following for each trade: The Call (Long or Short), the date the trade was placed, the price it was put on at, the date the trade was taken off, the price it was taken off at, the maximum draw down in pips (points), the draw down as a percentage of the opening price and final result (the win or loss in pips/points). An improvement I would suggest is tightening the stop losses and seeing what the results would be, I'd suggest about 5% for a long term time frame. I'll leave those improvements for the reader. Hope this is of some use to somebody. Thanks for reading.

• Home
• Technical Trading Systems
• Trading Articles
• Miscellaneous Articles
• Disclaimer
• Links
• Contact Me

• The characteristics of the best Forex trading systems
• Trading Strategies that work
• Donchian Channels, do they still work today?
• Strange Currencies II Trading System

• Mechanical Forex