Stop-Losses and Invalidation

The short answer

A stop-loss is not a budget. It is a statement about the market: if price reaches here, the reason I took this trade no longer applies.

The distinction sounds academic and is not. It is the difference between a stop that carries information and a stop placed at a number you picked because losing more than that would upset you.

Invalidation first

Before entering, answer one question: what would have to happen for this idea to be wrong?

If you are long because price is holding above a level that has repeatedly attracted buyers, then decisive trade beneath that level is your answer. If you are long because a trend of higher lows is intact, then a lower low is your answer. If you cannot state an answer, you do not have a trade — you have a direction you prefer.

That level is where the stop goes. Then, and only then, the position size formula tells you how many units keep the loss inside your risk limit.

Doing it in the other order — size first, stop wherever that size makes the loss tolerable — produces a stop that has nothing to do with the market. It sits at a price no participant cares about, and it gets hit by ordinary noise, and the trader concludes the market is out to get them. The market did nothing unusual. The stop was placed at an arbitrary point.

Why obvious levels get reached

Stops cluster. Just below an obvious support level, just above an obvious high, at round numbers — these are where a large number of participants place them, because these are where the chart looks meaningful to everyone at once.

A cluster of stops is a pool of resting orders. When triggered, they become market orders, and market orders are liquidity that large participants can transact against. Price tends to travel toward pools of liquidity, because that is where size can actually be filled.

You do not need a conspiracy to explain this, and there is no need to reach for the language of "stop hunting". It follows from the structure of a market: orders concentrate where the chart is legible, and volume goes where orders are.

The practical response is not to abandon stops. It is to place them where the idea is invalid rather than where the chart is obvious, and to accept that these are sometimes the same place. When they are, size down and widen — the arithmetic holds the risk constant.

Moving a stop

Three cases, with three different verdicts.

Widening it while losing. This is increasing risk after evidence has turned against you. It is the single most reliable way to convert a planned 1% loss into an unplanned 5% one. There is no version of this that is part of a plan.

Tightening it to break-even. Superficially free — the trade now "cannot lose". In practice you have moved the stop from a level that meant something to a level that means nothing except your entry price. Normal fluctuation now removes you from a position that has not been invalidated. Sometimes worth it, often not, and the deciding factor should be whether the market structure has genuinely changed, not whether you would like to stop feeling exposed.

Trailing it behind structure. Legitimate when defined in advance: as price makes higher lows, the stop follows beneath the most recent one. The rule must be written before entry, otherwise "trailing" becomes improvisation.

The common thread: a stop may move according to a rule you wrote before you were in the trade. It may not move because of how the trade feels while you are in it.

What a stop does not do

It does not cap your loss with certainty. A stop is a trigger that submits a market order, and a market order fills against whatever liquidity exists. In a gap, that can be far from your level.

This is why per-trade risk limits are set conservatively rather than at the maximum you could tolerate. The limit needs headroom for the occasional case where the stop does not do what it says on the label — and for the fees you pay on the way out, which the fee calculator will quantify for your own venue.