The thing that finally made stop losses click for me was noticing that most of the ones I set were getting hit by nothing. Not a reversal, not news, just the normal in-and-out breathing of the market. I would put a stop 2 percent below my entry, feel disciplined about it, and then watch a completely healthy trade knock me out on a random wick before doing exactly what I expected. The stop was not too tight in some abstract sense. It was too tight for how much that particular thing moves on a boring day, and I had no way of knowing that because I was picking the distance out of habit rather than measuring anything.
Average True Range fixes the measurement problem. ATR is just the average size of a bar's full range over some lookback, usually 14 periods, with gaps folded in so overnight jumps count. It answers one narrow question: how far does this asset typically travel in one bar right now. That number is the whole game, because a stop should sit outside the noise, and ATR is a decent proxy for how loud the noise is.
Why fixed-percentage stops misfire, especially on crypto
A fixed percentage stop pretends every asset has the same personality. A large-cap stock grinding sideways might have a daily range of well under 1 percent. A mid-cap token can swing 8 or 10 percent in a session and consider that a slow day. If you slap a 2 percent stop on both, you have set a stop that is comically wide for the stock and paper-thin for the token. The token version gets stopped constantly, and each time it does you convince yourself the trade was wrong when really your stop just measured nothing.
Crypto makes this worse for a couple of reasons. Volatility is not just higher, it is regime-dependent, so the same coin might triple its typical range going into a big move and then collapse back to a quarter of it during a dead stretch. A percentage you picked last month is stale. On top of that, thin books and leverage cascades produce wicks that spike far past the real trading range and snap back in seconds. Those wicks eat fixed stops for breakfast. An ATR-based stop at least scales with the current mess, so when volatility expands your stop widens with it instead of sitting there waiting to get clipped.
Step one: set the stop at a multiple of ATR
The stop distance is ATR times a multiplier. That is the entire formula.
Stop distance = ATR multiplier times current ATR.
If you are long, your stop sits at entry minus that distance. If you are short, entry plus that distance. Say ATR is 40 dollars and you are using a 2x multiplier, your stop is 80 dollars away from entry, wherever that lands in price. The multiplier is the one judgment call, and it is where most of the thinking goes.
Rough starting points I use, and these are starting points, not laws:
- Lower timeframes and scalps, tighter multiples, roughly 1x to 1.5x ATR. You are trying to be out fast and your ATR is already small.
- Swing trades on daily bars, roughly 2x to 3x ATR. You need room to sit through a normal pullback without getting shaken out.
- High-volatility assets and messy books, lean toward the higher end of whatever range you are in. A wicky altcoin at 2x will still get tagged more than a calm large-cap at 2x, so give it more.
The way to sanity-check a multiple is to pull up a chart and look at where a stop at that distance would have landed over the last few weeks of normal chop. If it would have been hit repeatedly by moves that went nowhere, it is too tight. If it sits so far out that a genuine trend reversal would cost you a fortune before triggering, it is too loose. You are hunting for the distance that ignores noise but respects an actual change of mind by the market.
Step two: derive position size from the stop, not the other way around
This is the part people skip, and it is the part that makes the whole thing coherent. Once you know your stop distance in price, you never guess at position size again. You decide how much of your account you are willing to lose if the stop hits, and you back into the size.
Pick a fixed risk per trade, something like 0.5 to 1 percent of account equity. Keep it constant across every trade. Then:
Position size = (account equity times risk percent) divided by stop distance per unit.
Concrete version. Say you have a 50,000 dollar account and you risk 1 percent, so 500 dollars is the most you will lose on this trade. Your ATR stop distance came out to 80 dollars per unit. Then 500 divided by 80 is 6.25 units. That is your size. If the stop gets hit, you lose right around 500 dollars, no matter what the asset is or how volatile it happens to be.
The quiet magic here is that every trade now risks the same slice of your account even though the stop distances are wildly different. A tight stop on a calm asset lets you hold a large position. A wide stop on a volatile one forces a small position. The volatility does the sizing for you. You stop having the situation where one oversized altcoin trade can do more damage than five stock trades combined, because the math automatically shrinks the position when the thing is dangerous.
The failure modes worth knowing
A few things that will bite you if you are not watching for them. First, recompute ATR when volatility regime shifts. If you sized a position off last week's calm ATR and the market wakes up, your live risk is now larger than you think because price can travel your stop distance and then some in a single bar. Second, leverage does not change the risk math but it absolutely changes whether you get liquidated before your stop triggers, so on a perp make sure your ATR stop sits well inside your liquidation price or the exchange decides your exit for you. Third, resist the urge to move a stop that is doing its job. If your ATR distance was honest, a stop getting hit is information, not an insult.
When I am watching a lot of markets at once, I lean on tooling to keep the ATR and the scorecard current across assets so I am not eyeballing volatility by hand, which is roughly the reason we built the market scorecards into Blockcircle. The formula is simple enough to run on a napkin, though. Measure the noise, put the stop past it, size the position so the stop costs a fixed percentage. Do that consistently and your worst trade stops being able to hurt you more than any other one.