Conceptually, robust trend following systems are designed around the basic principles of identifying and capturing directional price movement, without having layers of complexity or being tailored to individual markets.
This leads to volatility in the performance achieved, as market conditions move through phases of 'trendiness' and differing levels of price volatility, but that is the compromise to robustness.
So when people suffer a period of poor or non-performance and start questioning the viability of trend following, what they are actually saying is that their own systems are currently unable to profit from the price trends being generated by the market.
As Ed Seykota told us, there is no such thing as THE trend.
So while price trends are still there, they are simply occurring with differing levels of magnitude or volatility, or over a longer or shorter timeframe, compared to what their systems are designed to capture.
As a result of this, the mindset I have developed is to view my entry / exit rules simply as a mechanism to interact with a market.
Do the rules or parameters that I use have any magic qualities?
No, of course not.
I accept they won't capture a lot of price trends. But that's not what they are designed to do.
What they DO give me is a framework allowing me to get into a potential new trend (as determined by my timeframe and parameters) and get me out either if that trend doesn't develop or when that trend has run its course.
Nothing more, nothing less.