One way to handle unpredictable ever-changing price sequences is implement a non-curve fit entry/exit rule. The point of the rule is not to make money all the time in every market, but to keep you in business and allow you to exploit trends from time to time.
In systematic trading, much more important than entry/exit parameters are the markets you trade and don’t trade; then position sizing.
Sticking with the SPY example, if your system buys on a simple moving average crossover of 50/200 then it’ll probably do pretty well from 2009–2015. It probably won’t perform well near the top in 2007 or the recent slide this past August. But the point is not to earn profits all the time.
When the entry/exit indicator doesn’t perform well in any one particular market, that’s when portfolio diversification and position sizing comes in to keep those losses small and to possibly move capital from the losing positions to maybe some other market that’s trending well at the moment.