A trailing stop is a stop-loss that automatically adjusts upward (for a long position) as price moves in your favor. It can be defined in dollars (trail $0.50 below the high), percent (trail 2% below the high), or by structure (trail to the prior swing low, or below a moving average).
The trade-off is mechanical: a tighter trail locks in more of the current paper gain but cuts winners short on normal pullbacks; a looser trail captures more of the eventual continuation but gives back more profit when the move finally reverses. There is no setting that wins on both.
Trailing stops convert the binary stop-loss into a dynamic exit. Initial stop defines the maximum loss; trailing stop converts the trade into a winner-of-some-size once price has moved enough in your favor. Most journal-driven strategies use both — a hard initial stop until price is at +1R or so, then a trailing stop to manage the runner.
Manual trailing is where most traders lose discipline. Moving the stop tighter after every bar feels productive but invites confirmation bias and tilt — you end up exiting on noise rather than structure. Rule-based trailing (trail below the prior bar's low, or trail to the 20-EMA) outperforms ad-hoc trailing in nearly every journal I've reviewed, because it removes the in-the-moment temptation to micromanage.