I agree but even then with discretion, based on price action. It can stop you out even though the price is going your way, just because it ratchets down over time. I've also used the Linear Regression in a similar way, but that too can do the same thing. While we're on that subject, what is the/a good way to use the Linear Regression. I've been using it to measure entries/ exits based on strength/weakness at Support and Resistance levels, but can be quite quirky.
Any ideas?
That depends on what you use it for, how you use it and what your definition of reliable is. Price by itself is unreliable. At any given point in time it can do anything. A moving average crossover is unreliable by its self and means very little but at a support or resistance level it can mean a great deal.