What a stop order is

The instruction itself 

A stop order is a level with an instruction attached to it: buy, or sell, but only once the market reaches the level. Until then nothing happens. The order rests where it was written, it transacts nothing, it appears in a pending list rather than in a position list, and it has no price of its own in any meaningful sense. When the market reaches the level, or passes it, the instruction activates and is submitted as a market order, which means it transacts at the best price then available rather than at any price it named. The level that activates it is called the trigger price, and it is the only price the order specifies. Everything that happens after the trigger belongs to the market.

Two events, not one 

The limit order in the previous lesson names a price and waits. It will not transact at a price worse than the one named, and the cost of that certainty is that it may never transact at all. A stop order inverts both halves of that arrangement. It names a level that decides when it transacts and accepts whatever price is available at that moment, so the cost of near certain execution is uncertainty about the price received. A limit order fixes price and leaves timing open. A stop order fixes timing and leaves price open. Reading the two side by side is the quickest way to keep them apart, and mixing them up is the most consequential error in this module, because an instruction written as a stop where a limit was meant transacts at a price nobody sanctioned.

Which price triggers it 

A quote is two prices, and a stop reads only one of them. A sell instruction transacts at the bid, so a sell stop is conventionally triggered when the bid reaches the level. A buy instruction transacts at the ask, so a buy stop is triggered when the ask reaches it. The asymmetry matters most for a level written a short distance from the current quote, because the price displayed large on most screens is only one of the two: a level that still looks untouched on a chart drawn from the bid may already have been reached on the ask. Which side each order type reads is stated in the contract specifications and in the platform's own documentation, and the convention is not identical across every venue, which is why it is published per venue rather than treated as universal.

Entry stops and protective stops 

Mechanically there is one stop order. Conventionally there are two names for it, and what separates them is intent rather than machinery. The geometry is fixed by the definition: a buy stop sits above the current market, because an instruction to buy that is not wanted yet can only be waiting for the market to rise into it, and a sell stop sits below the current market for the mirror reason. Nothing in the order records why it was written. The same instruction opens a position for one trader and closes one for another, and the platform matches it identically in both cases.

• A buy stop above the market, with no position open, opens a long if the market rises to the level. Traders call this an entry stop, or a breakout entry.
• A sell stop below the market, with no position open, opens a short if the market falls to the level. It is the mirror of the first case and carries the same names.
• A sell stop below the market, with a long position already open, closes that long if the market falls to the level. Traders call this a protective stop, or a stop loss.
• A buy stop above the market, with a short position already open, closes that short if the market rises to the level. Same instruction, opposite direction.

One bookkeeping difference does follow from the intent, and it is worth knowing because platforms handle it differently. A protective stop is usually recorded as an attribute of the position it is attached to, so closing that position by any other route cancels the stop with it. An entry stop stands alone, and once it fills it leaves an open position with no protective instruction of its own unless a separate one was written alongside it. Traders who work primarily with entry stops meet that gap early: the order that created the position had nothing to do with ending it, and the two instructions are related only by the sequence in which they were written.

The distance between the trigger and the fill 

The difference between the trigger price and the price the resulting market order transacts at has a name. It is not a fee, it is charged by nobody, and it appears on no statement as a line of its own. It is the arithmetic consequence of the market moving in the interval between the level being reached and the market order transacting, and of the size of that order relative to the prices available at the instant it arrives. It runs in both directions. A fill can land beyond the trigger, and it can land better than the trigger when the price recoils in the same interval.

When the market gaps through the level 

A difference of a few increments is ordinary. A gap is the extreme case, and it is a different phenomenon rather than a larger version of the same one. A gap occurs when two consecutive trades happen at prices that are not adjacent: a market closes on a Friday and reopens somewhere else entirely, a scheduled release reprices an instrument between one quote and the next, a thin market skips several increments in a single move. Where a gap spans the trigger, no trade happened at the level at all. There was nothing there to transact against. The order is activated by the first price that reaches or passes the level, and it fills at the best price then available, which sits on the far side of the gap.

Reading the implied distance and the realised distance together is the point of the block. The level is an instruction about when, and the gap is a statement about what the market was willing to trade at. Nothing inside the order can bridge the two, which is the mechanical reason a stop order cannot honestly be described as putting a floor under a loss: it fixes the point at which an instruction is sent, and nothing whatever about the price that instruction receives. A separately priced order type that does guarantee the level exists at some venues, usually under a name such as guaranteed stop; where it exists it is a distinct product with its own conditions and its own charge, and whether any given firm offers it is a matter for that firm's contract specifications rather than something the word stop implies on its own.

Where practitioners disagree 

Two arguments about stop orders are genuinely unsettled. The first is about clustering. One tradition holds that resting stops accumulate just beyond round numbers and recent extremes, that the accumulation can be inferred by participants who see enough of the market, and that price is therefore drawn toward those levels before continuing. Another answers that the claim is close to unfalsifiable as usually stated, since any level that is reached can be described afterwards as having attracted the market, and that moving a level somewhere less obvious only relocates it to whatever the next obvious place is. Both traditions concede the one part that is arithmetic rather than interpretation: a level written close to the market is reached often and a level written far from it is reached rarely, and that is a fact about distance, not about anybody's intentions.

The second argument is whether the instruction belongs on the platform at all. A resting order acts without attention, including in hours when nobody is watching the screen. Holding the level in mind instead, which practitioners call a mental stop, leaves no instruction anywhere, and the tradition that prefers it argues that a resting order transacts on a momentary spike that reverses seconds later, converting a temporary move into a closed position. The tradition that opposes it observes that a mental level does not exist during a gap, cannot act while the holder is asleep, and depends on a decision being taken at precisely the moment when taking it is hardest. Neither position removes the trade off. They disagree about which of the two failures is worse, and the honest answer differs by instrument, by the hours a market trades and by how much it tends to move while unattended.