A Donchian channel is a market indicator developed by Richard Donchian; the channel is typically formed by taking the highest high and the lowest low the market in question has made in the previous n number days – when the price moves outside of the channel it is called a breakout. A common variation of the standard Donchian channel is a channel made from the highest close and lowest close in the previous n number of days with a breakout defined as a new higher or lower closing price.
Donchian channels are typically used in technical trading systems and tend to work best on large liquid markets that trend well, most notably the EUR/USD Forex currency pair. Donchian channels were a key component in the trading system used by the Original Turtles in Richard Dennis and William Eckhardt's famous experiment in the 1980's.
But do Donchian channels work today? As is the answer to many questions like this, I suspect the answer is it depends. Donchian channels are only one component (albeit a key one) in the technical trading systems that use them and whether the system using them will work will depend upon the markets on which they are used, how they are used, and the other components of the trading system (e.g. exits, stop-losses, pyramiding strategy, etc...) that they are used with. However, one thing that does appear to have changed since the days of the Original Turtles is the time frames that the breakouts are effective on. William and Richard's system used a 20-day (4 week) Donchian breakout as an entry with a 10-day (2 week) Donchian breakout in the opposite direction as an exit signal, or a 55-day breakout as an entry signal and a 20-day breakout in the opposite direction as an exit signal. These time frames are not nearly as effective today.
I am going to test how effective Donchian channel closing price breakouts have been fairly recently on the EUR/USD currency pair and the S&P 500 stock market index. The Donchian entries will be tested with Donchian exits of half the entry size from the beginning of the year 2000 to the end of 2009. Here are the results –
The markets it is used on: As I thought, Donchian channels work far better on currency pairs (particularly the EUR/USD) than they do on stock market indices. With stocks it's generally a good idea to look to buy low and sell high. However with currencies it's actually a far better plan to aim to buy high and sell even higher; or to sell low and buy back even lower. Currency pairs don't have a directional bias and the Forex market is very large. Once a trend gets underway on a large liquid market such as the EUR/USD it actually takes an awful lot to stop it, when a trend has begun it is more likely to continue than not. As a general rule the bigger the market and the stronger the trend the greater the chance is that it will continue rather than reverse.
The exit it is used with: Using a time-based exit as opposed to a Donchian channel exit tended to perform slightly worse on the market where Donchian entries performed well, and slightly better on the market where Donchian channel entries performed badly, at least it did in terms of the number of pips won to the number of pips lost ratio (which is by far the most important result). This tells me that using Donchian channels as exit strategies also have a slight edge on markets where Donchian channels tend to work well. However, in terms of the percentage of winning trades, time-based exits generally produce more winning trades than losing trades whenever the system's entry criteria has an edge. On the EUR/USD only a 20-day entry didn't have an edge, probably because it is so commonly used that it has ceased to be effective.
The time frame: As a general rule Donchian channels work well on large time-frames, the larger the time frame the more likely the breakout is to be significant.
