I'm wondering about methods for determining when a system no longer works. I've read people suggesting you should ride a system until it stops working, then jump onto another system that's currently working - and focus less on long term historical performance (I'm not advocating this).
What methods are there for determing is a system is working - obviously if there's a really smooth equity curve, you can look for deviation form that curve. But I wonder if you might have to give back your gains before this becomes obvious, in the real world. Are there other methods to consider? All conributiosn appreciated, Rich
