In general, proper trading strategies exploit certain broad "inefficiencies" of a market (e.g. trends, mean-reversion, calendar, etc).

Such inefficiencies are "probabilistic", and so performance is significantly a function of market conditions. There will be periods when the strategy is MOSTLY in "sync" with the market and periods when it is not.

With smart position sizing and/or equity curve trading you can normally improve performance, but you still need to trade/track the strategy continuously.

And, importantly, also monitor the "inefficiency"/your algo itself. That's what "Cold Blood index" is for.

The Zs are "fixed" in terms of the inefficiencies they exploit.
But:
1) the algos themselves are occasionally improved (or retired), and
2) they are regularly "re-trained", i.e. parameters used are recalibrated to the current market conditions.