Why does price keep hitting your stop loss?

It is one of the most demoralising experiences in trading. You are right about the direction. Price goes against you just far enough to take you out, then turns around and runs exactly where you thought it would. It happens often enough that it stops feeling like chance.

The usual explanation is that your broker is hunting you personally. That is almost never what is happening, and believing it will keep you making the same mistake. The real reasons are duller and far more fixable.

See it on a real chart: Gold (XAUUSD) liquidation heatmap — the modeled clusters above and below the current price.

Reason one: your stop is sized to your wallet, not to the instrument

Most traders decide how much they are willing to lose, convert that to a distance, and place the stop there. That distance has nothing to do with how far the instrument routinely moves. If gold swings 12 dollars in an average hour and your stop sits 8 dollars away, you have not set a risk limit — you have bought a lottery ticket on the next hour being calmer than usual.

The fix is to start from the instrument. Look at its average true range on your timeframe, place the stop beyond normal movement, and then size the position so that distance costs you an acceptable amount. Distance first, size second. Most people do it in the opposite order, which is why they get stopped out by ordinary noise.

We measured how often that happens. Treating every hourly close over two years as an entry, long and short, a stop one hourly ATR away was touched within 24 hours 72% of the time on gold, 76% on EUR/USD, 74% on Bitcoin and 77% on the Nasdaq-100. At two ATR it was still close to half. In other words, a stop inside one ATR is hit on most days whether or not the trade idea was right.

The full table for each market, with today's distances in dollars, pips or points, is on the gold stop-loss data page and its equivalents for forex, crypto and indices.

To run that sequence on gold without doing the maths by hand, place entry and stop on the chart of the free gold stop-loss and position size calculator. It shows how large a position the stop allows for the risk you set, and how far the stop sits in multiples of the current hourly ATR.

Is it because your stop is where everyone else's is?

Ask a hundred traders where to put a stop on a long and most will say the same thing: just below the recent swing low. Below yesterday's low. Under the round number. These are reasonable answers, which is precisely the problem — they are reasonable to everybody, so the orders pile up in the same few places.

A cluster of resting stop orders is a pool of liquidity. Someone who needs to fill a large position has to find a counterparty, and a pocket of orders that will fire automatically is a convenient place to look. You do not need a conspiracy for this; it follows from the fact that large orders need liquidity and stops are liquidity.

Does that make a stop on the obvious level more likely to be hit? In our tests, no: a stop on a swing low or round number was hit no more often than one the same distance away. The pocket tells you whose company your stop is in; distance and the hour you enter decide how often it is hit.

Reason two: the spread reaches further than the price you see

Your stop triggers on the broker's quote, not on the last traded price on a chart. When volatility spikes — a data release, a session open, thin liquidity — the spread widens, and the effective price that reaches your stop can be meaningfully further than the candle low suggests. This is why your stop gets taken by a wick that, on the chart, appears not to have touched it.

Ask your broker what the typical and maximum spread is on the instruments you trade, and add that buffer to your stop distance. On gold and index CFDs it is usually larger than people assume.

What actually fixes it

• Set stop distance from the instrument's volatility first, then size the position to fit your risk — never the reverse.
• Mind the hour as well as the distance: the same stop is two to four times as likely to be hit when opened just before the New York session as in the London evening (on gold, a 1 ATR stop: roughly 60% against 20-25% within 4 hours). Moving it just past the obvious shelf helps only by the distance it adds: in our tests a stop on the shelf was hit no more often than one the same distance away.
• Add a buffer for spread widening, especially around session opens and scheduled data.
• Before entering, you can look at where the modeled stop concentration sits above and below price — as context, not as places to target or to avoid: a stop on one was hit no more often than a stop the same distance away.

Where a stop-loss heatmap helps — and where it does not

A heatmap shows where a model estimates stop orders are concentrated, so you can see whether your intended stop is sitting in the crowd. That is the context in the last point above; it does not change how often your stop is hit.

It is worth being clear about the limit, because most of this industry is not. We backtested our own levels against a control — an arbitrary price the same distance away — across roughly 100,000 historical levels. Price did not reach our levels more often than the control, and did not reverse at them more often either. The levels describe where the model places concentration. They do not tell you which way price is going, and anyone claiming otherwise about their own levels has probably not checked.

That is the honest use: not a signal to trade, and not a list of places to keep your stop away from, but a map of where the model puts the crowd's stops. You can look at where the clusters sit on gold, forex, indices and equities in the browser, free on a 24-hour delay, with no card.