Do you have the rules used in the backtest? We backtested the turtle method on System Writer Plus early 1989(now you know I've been around the block a few times). The backtest results were exceptional. I bet hundreds of others were doing the same at the same time because when we started trading it late 1989 we made exceptional returns but suddently the system inverted completely and we lost most of our profits during the next two years. Since then it has never worked, I mean that particular system. Obviously, these turtle fund manager use something else.
The tradingblox forum has a few threads dealing with this topic in a lot of depth. The current thinking is that the Original Turtle System was tuned for a highly inflationary environment. The breakout periods (20 days for system 1, 55 days for system 2) were fairly short-fused, but got you into these big trends quickly. Most turtles added positions onto winning trades - which results in huge profits in large trends, but adds to overall risk. As inflation cooled, the trends got choppier and noisier, and whipsaws killed the system. Also, there were a lot of people who figured out the system and either front-runned it or faded it (Google Linda Bradford Raschke "Turtle Soup"). Go to the tradingblox forum and do a search under "dead turtles" and/or "robusti family" about different approaches to continually improve a trend following system.
True. But some of these folks looking at entry only, over some fixed portfolio that never changes, will never understand your logic.
This is simply not accurate analysis. It's almost as if you don't even know the turtles are just trend followers.
I would not trust anyone in here (including me). Go put rules, Turtle or whatever trend following system, into code and test them. See the nuances. Clearly, trend following has not stopped working, and yes basic robust systems still work, but the real success keys start with portfolio selection and money management. Then the system.
The founder/CEO of the leading London-based trendfollowing firm covered this and other topics in a nice presentation in NYC earlier this week. I didn't see Mike in the crowd, though he may have been -- I didn't scan closely. If the firm posts the slide-deck online, it would be worth a look.
One part of the system that I do not understand is that in order to take a signal, the previous signal must have failed. What exactly does that mean and why was that a rule?
"The turtles ignored the System One four-week breakout signal if the last four-week signal was a winner" -from Covel's book.
So Covell, how many "courses" (not books) did you sell as the result of this thread? ...just curious, not offence...