The Trouble With Turtles

I've been using the Turtle trading method for much of this year to identify trades. During my August recess, I spent some time thinking about the results.

Turtle trading involves playing breakouts beyond the 55-day Donchian price channels, which mark the high and the low of the last 55 days.

It's all hedged around with other rules (see my essay "How I Trade"), but essentially, it's a price-level breakout scheme.

It has the advantage of being well defined, and of showing up clearly on a chart. Those qualities allow me to scan 600+ stocks in an hour, looking for viable trades.

Another advantage is that the method is very selective. It doesn't produce a lot of trades, allowing the turtle trader to focus on those opportunities that are, presumably, the very best.

The method's use of the average true range in setting stops, thereby accounting for volatility, is one of the more useful concepts I've encountered.

Yet, there are troubles with Turtles.

• The 55-day price channel is entirely arbitrary, unrelated to the actual congestion points in the price movement. This has two implications:
• Lack of confidence. If 55 is just a magic number, then why not use 40, or 20, or 77? As a fact-based trader, I need to know that my method has reason behind it.
• Ill-defined scalability. If 55-day channels are used on the daily chart, then what number should be used intra-day, on a half-hour chart or a five-minute chart? Nobody knows, since the original 55-day standard is arbitrary.
• The adjunct rules are also arbitrary. For example, the two-day rule, which requires an exit from a position if the price closes within the price channels for two consecutive days after breakout. What's so special about two days? Nothing. It's arbitrary.
• The method is too selective for some markets. It produces a fair number of trades when it comes to individual stocks, but fewer for exchange-traded funds, and far fewer for currencies. During my August vacation, I limited my currency scans to the 14 most liquid pairs, and there were hardly  any breakouts for the entire month. I can't make money if I don't trade. 
• In many instances,  the breakout level is in fact a price reversal level. The USD/JPY (U.S. dollar/ Japanese yen) currency pair, in particular, seems to be prone to reversal right at the 55-day price channel boundary. Could it be that the Bank of Japan is filled with Turtle traders?
• The original Turtle rules can produce huge losses. Under those rules, entry is breakout beyond the 55-day channel, and exit is breakout in the opposite direction beyond the 20-day channel. That can be a very wide spread. For a leveraged position, especially, the losses can be crippling.

That's my critique of the Turtle trading method. I'm looking at other breakout schemes, such as that outlined by Joe Ross in his classic, Trading by the Book.

I'll post on the subject I get the rules worked out. (Ross is a great trader but is fairly disorganized a writer. It takes a lot of work to put his method into a succinct set of rules.)