Trailing stop

A stop loss order whose level is recalculated as the market moves. The instruction names a distance rather than a price. On a long position the level sits that distance below the highest price reached since the order was set, and it is moved up whenever a new high is made; it is never moved down. On a short position everything is reversed. The effect is a level that ratchets: it locks in the distance travelled and does not give it back.

How the distance is set divides practice. A fixed number of price increments is the simplest form and is what most platforms implement. A distance scaled to recent volatility, commonly a multiple of the average true range, adjusts as conditions change and is favoured by those who observe that a fixed distance is too tight in an active market and too loose in a quiet one. There is a third convention that trails behind a moving structure such as successive swing lows rather than behind the price itself. None of the three is established as better, and the trade off between them is the same in every case: a closer distance ends more positions early, a wider one returns more of the movement before it acts.

Two limits are frequently missed, and they are limits of the mechanism rather than of any implementation. Once triggered, a trailing stop becomes an order to deal at the next available price, so a gap or a fast market fills it some distance from the level it named: the level is where the instruction activates, not a price it secures. And the trail follows price, which is not the same as following the direction of the move, so a single sharp movement against an otherwise intact position can reach the trailing level and end the position, an outcome traders describe as being taken out on noise.