What it is
Managing risk means deciding before you enter how much you accept to lose if you are wrong, and getting out when that happens. It has three pieces that go together: where you put the stop, how big the position is and how much leverage you use.
None of the three guesses where price will go. They do something just as important: when the trade goes wrong, which happens often, the loss is small and it is one you chose.
The stop: where to put it
The stop goes where the trade idea stops making sense. If you buy because price reaches a support, the idea is invalidated if price breaks through it: the stop goes a little beyond the zone, not right on it, because price often pokes through levels before carrying on.
How far beyond? We measured it. We took every arrival at a level that holds on 1-hour BTC, ETH and SOL candles since 2021 (about 24,000), entered in favour of the textbook reading and put the stop behind the zone with four different margins, measured in ATR (the average hourly volatility). The target, always twice the risk, with 48 hours to get there:
| Margin past the zone | Typical distance | Reaches target | Stops you out, then gets there | No level |
|---|---|---|---|---|
| Tight (no margin) | 0.75% | 31.2% | 55.6% | 57.4% |
| ¼ ATR | 0.96% | 31.9% | 45.8% | 49.4% |
| ½ ATR | 1.21% | 31.8% | 36.4% | 41.5% |
| 1 ATR | 1.69% | 31.5% | 22.0% | 30.3% |
How to read it:
- With the stop right on the level, half the time it takes you out and then price proves you right. But it happens almost as often at any random hour, with no level (57.4%): nobody is hunting your stop, it is the normal back-and-forth of price.
- Giving it room cuts those sweeps a lot: with a 1 ATR margin they drop to 22%.
- The target is reached just as often with all four (31–32%, ±0.6). With a target at twice the risk, with no edge at all you would expect 33%: the stop brings no edge by itself; it brings a plan.
- More room means more distance: to risk the same, the position has to be smaller (see below).
Move the stop?
It is the most human mistake: price gets close to the stop and you move it away "to give it room". We measured it in the 15,686 cases of the study where price touched the stop (½ ATR margin), comparing respecting it with moving it away up to twice, as someone who does not want to take the loss would:
| If you move it… | Of the cases |
|---|---|
| …it reaches the target (+2 times the risk) | 27.5% |
| …you end up losing three times the plan | 48.7% |
| …it is still open after 48 hours | 23.8% |
Respecting it, you always lose what you planned. Moving it, sometimes you get away with it, and that is what makes it so tempting; but on average you lose 4.7% more (±3.4), and almost half the time you lose three times as much. The market is not about being right: price decides.
Position size
The stop says where you get out; the size says what getting out costs you. The rule is a division:
Size (as % of the account) = what you accept to lose (as % of the account) ÷ distance to the stop (in %)
| Stop | Distance | Position size |
|---|---|---|
| Right on the level | 0.75% | 133% of the account (needs a little leverage) |
| ½ ATR | 1.21% | 83% of the account |
| 1 ATR | 1.69% | 59% of the account |
Notice what it says: with a well-placed stop, leverage is only needed to fine-tune the size, and very little of it. It is not there to win more, but to risk exactly what you decided.
Leverage and liquidation
With leverage, if price moves far enough against you, the exchange closes the position and you lose the whole margin: that is liquidation. In a simple calculation, the distance to liquidation is roughly 1 ÷ the leverage. On a real exchange it comes sooner, because of fees and maintenance margin.
| Leverage | Liquidated by | BTC | ETH | SOL |
|---|---|---|---|---|
| 5x | 20% | 0.02% | 0.08% | 0.45% |
| 10x | 10% | 0.47% | 1.13% | 3.42% |
| 20x | 5% | 4.03% | 7.77% | 15.95% |
| 25x | 4% | 6.82% | 12.39% | 22.70% |
| 50x | 2% | 24.03% | 34.11% | 48.38% |
At 50x, an ordinary move liquidates you in BTC: it happens 1 in 4 times within half a day, and 1 in 2 in SOL. At 5x, it takes a move that happened 1 in 4,585 times in BTC. The practical rule: the stop must always sit well before liquidation. If liquidation comes before your stop, you have no stop.