Where a Stop Belongs

A stop has one job: to close the trade at the point where the reason for taking it has stopped being true. It is not a device for limiting losses to an amount you like the look of. That job belongs to the position size, and confusing the two is how traders end up with a stop that is both too tight to survive and too large to afford.

Invalidation, defined before entry

Before the order, finish this sentence: "I am wrong about this trade if price does this …" Whatever fills the blank is where the stop goes.

• "...closes back below the range it just broke out of" — the breakout failed.
• "...trades below the swing low the bounce came from" — the level did not hold.
• "...moves more than a day's typical range against me" — this was not the move I thought it was.

Each of these is a statement about the market. None of them mentions your balance, and that is the test. If the sentence you wrote contains a dollar figure, you have written a sizing rule and called it a stop.

Structure stops

The most common approach puts the stop just beyond a level the chart has already respected — the low a bounce started from, the high a rejection came off, the edge of the range that just broke.

The logic is that price trading through that level means the participants who defended it are no longer defending it, which is a different market from the one you entered.

Just beyond, not exactly on

Put the stop at the level and you will be taken out by the wick that tests it. Every obvious level attracts a cluster of stops, and price has a persistent habit of reaching into that cluster before continuing in the original direction.

A small buffer beyond the level — a few pips on a major, more on a volatile pair — is the usual answer. It costs a little on every stop-out and saves the whole trade on the ones that would otherwise have been shaken out for nothing.

Volatility stops

The alternative sets the distance from how much the instrument typically moves, rather than from a level. The usual measure is Average True Range — the average distance covered per bar over the last N bars, normally 14.

A stop of one to two ATR beyond entry is the standard construction, and it has two properties structure stops do not:

• It adapts automatically. When a pair gets more volatile, the stop widens and the size shrinks, without you deciding anything.
• It is comparable across instruments. Two ATR on GBP/JPY and two ATR on EUR/USD represent the same amount of ordinary movement, even though one is four times as many pips.

What it lacks is a reason. A volatility stop says the move is larger than usual; it does not say the idea was wrong. Many traders use both — the structural level as the logic, ATR as a sanity check that the level is not absurdly close or far in the current conditions.

What to do when the stop is too wide

Here is the situation that produces most of the damage. The level that invalidates the trade is 90 pips away. Sized properly on a $2,000 account risking 1%, that is 0.02 lots — a position so small it feels pointless.

There are four honest responses and one dishonest one.

1. Take the small position. It is correctly sized. Feeling pointless is not an argument, and the trade returns the same number of R as a larger one would.
2. Find a closer entry. The stop is fixed by the chart, but the entry is not. Waiting for a pullback closer to the level shortens the distance without moving the invalidation.
3. Drop to a lower timeframe, where structure is tighter — accepting that the levels are correspondingly less significant.
4. Skip it. A trade you cannot size properly is a trade that does not fit your account today. There will be another.

And once it is placed

Widening a stop because price is approaching it is the same mistake made later and larger. The level was chosen when you were calm; the urge to move it arrives precisely when you are not.

Moving a stop closer — to break-even, or trailing it behind a trend — is a different action and a legitimate one, provided the rule for doing so was written before the trade rather than invented during it.