"We have a saying here: "It is incredible how rich you can get by not being perfect." We are not looking for the optimum method; we are looking for the hardiest method. Anyone can sit down and devise a perfect system for the past." - Larry Hite
As a trend follower, it is important to acknowledge and accept that individual stocks making up the 'stock market' are in a constant state of change themselves, be it trending or non-trending, and all with differing levels of volatility, as well as the indices themselves, along with foreign exchange, interest rates, commodities etc.
Therefore we always need to be wary of the siren call of backtesting, and whether you should amend your approach based on recent past performance, when it is a given that, providing markets do not operate under price controls, they will continue to be changing their state.
Recently we have seen that Winton Capital has decided to reduce its exposure to trend following following a prolonged period of poor performance, combined with the belief that the future performance of such an approach will continue to erode.
But isn't that, as Ed Seykota would say, an example of trying to predict a non-existent future?
From a trend following perspective, one thing I have found out over the years is that simplicity beats complexity, and also assists with the overall robustness of your approach.
As Richard Dennis said in his Market Wizards interview, a 'robust' method should be able to generate profits "whether you are trading bonds or beans".
Yes, you will still suffer periods of poor or non-performance (usually when price movements quickly reverse and/or when volatility increases), which is why you need strong risk control.
But providing there are no restrictions on price movements, trends will continue to appear and be available to profit from. In that sense, a simple and robust trend following method will remain able to capitalise from those moves, without the need to constantly fine-tune the parameters.