Disliked{quote} {quote} lol, hope your testing goes well. My gripe with filters is the implicit suggestion that they add value with no drawback. People tend to think of filters in the context of "filtering out bad trades". After having spent much time wrestling with these ideas I came to the conclusion that simplicity prevails for two reasons: 1. More filters increase the likelihood that a good historical performance is the result of curve-fitting. 2. The inescapable fact of trend trading is that profits come from being exposed to market...Ignored
I previously developed a trend following strategy for trading equities. By the introduction of one 'key' filter the bad trades did in fact reduce, as well as overall drawdown to the system. This was then back tested over 10 years of clean historical data ensuring that each block of new data backtested was out of sample. I even ran monte carlo analysis on the backtested data and found the overall profit & max DD distributions to be extremely tight.
The lesson learn (for me anyway) is that they shouldn't be discounted but there does need to be real logic behind why you think they might work.
Just my 2cents.
CCH