I blame the Japanese, taught English (Australian slang) over there for 2 years and a "coke high" was the drink de jour at the time, I even despoiled the Chivas Regal I found in my apartment, and that's a pretty sweet Scotch to start with
Has anyone done a back test (using at least 6 months of past data) on the original method? I was hoping to find the results of such a test somewhere in the contents of the first 100 posts, but I gave up after reading the first 500 posts. If that has not been done, I'd be happy to be the first to do it. However, it's very hard to fish out all the bits and pieces of the system out of this very long thread. Is there a summary with all the rules well defined so that they can be directly coded for a back test? It's entertaining to look at the 1-day charts to see how this works (or not), but for a sound evaluation, a much longer test interval (and unambiguous entry/exit rules) are needed.
I can promise you that it will not test out well. That doesnt mean it wont work. Discretion is a large component.
I amended the post to include discretion. I do not see how backtesting any system based on a lagging indicator could show a positive result over the long run.
I am still confused. Why does it matter if it's a lagging indicator? The point of a back test is to see if the system has a positive expectancy over a reasonably long period of time. Are you saying that the back test for any system that uses lagged indicator has no validity, or perhaps that the particular system discussed here is not really profitable?
Go ahead and test it. The results will not be good. Trader28 says he has been trading this profitably for a couple of years. It boils down to: are you going to believe the backtest or Trader28? Most quant types will believe the backtest and they will never make any money trading it. Backtesting cannot factor in discretion. edit: before you ats traders jump on me notice I said "most". I remember reading an article by linda raschke where she commented that after 20+ years of system trading she only knew of 2 people who made money trading a totally automated system so it is possible.
What bb is tryiing to say is that this system is a guideline so a backtest wouldnt work because the system IMO is more then 75% the person useing it.
But in the back test, there is no "person", so it's very objective: it simply follows the rules of the system. So, I am still not sure what you guys are talking about. If the system in question requires "discretion" (which, I suppose, means going against the system from time to time), then this "system" is not really a system, is it?