Risk management in crypto trading: position size, stop-loss and R-multiples

Most trading accounts are not destroyed by a bad idea. They are destroyed by size: a position too large for its stop, a stop moved when it was about to be hit, a leverage setting chosen because the exchange allowed it. The idea behind the trade can be perfectly reasonable and the account still ends at zero.

Risk management is the set of decisions that keeps a single trade — or a single bad week — from ending the game. This article walks through them in the order they happen: risk per trade, position size, leverage and liquidation, R-multiples, stop placement, fees and funding, drawdowns, and several positions at once.

Every number below comes out of the formula shown next to it, so you can redo it for your own account. Nothing here is investment advice; it is arithmetic and a few widely used rules of thumb.

Risk per trade: decide the loss before the entry

The first decision is not the entry price. It is the amount you are prepared to lose if the trade is wrong. Decided before the entry, it is a calm number; decided after, while the position is open and red, it tends to grow.

A common rule is to risk 1–2% of the account per trade. The percentage matters more than the exact figure: as a share of the balance, the risk shrinks automatically after losses and grows after gains, which stops a losing streak from compounding. Ten consecutive 1% losses leave about 90% of the account; ten 10% losses leave about 35%.

The alternative is a fixed amount in USDT — say 100 per trade — which some traders prefer because it is easier to think in and does not creep upward after a good month. Either way, the loss is written down before the order; everything that follows derives from it.

Position size comes from the stop, not from leverage

Once the risk amount is fixed, the position size is no longer a matter of opinion. The distance from the entry to the stop-loss tells you how much one coin can lose; divide the risk amount by that distance and you get the number of coins for which hitting the stop costs exactly the planned amount. Leverage appears only at the last step, when the exchange decides how much margin to lock.

Worked example: balance 10,000 USDT, risk 1% = 100 USDT. You want to long BTC at 100,000 with a stop at 98,000. The stop is 2,000 away (2%), so the quantity is 100 ÷ 2,000 = 0.05 BTC. The position is worth 0.05 × 100,000 = 5,000 USDT, and at 5× leverage the margin is 5,000 ÷ 5 = 1,000 USDT. If the stop is hit you lose 0.05 × 2,000 = 100 USDT — the 1% you decided on.

Add a take-profit at 106,000 and the reward is 0.05 × 6,000 = 300 USDT: 6,000 of distance against 2,000 of risk, a reward-to-risk of 1:3. The position size calculator does this arithmetic for any balance, risk and stop.

Notice what did not enter the calculation: leverage. The same 0.05 BTC position needs 5,000 USDT of margin at 1×, 1,000 at 5× and 500 at 10×, and loses the same 100 USDT at the stop in every case.

Leverage changes the margin, not the risk — until liquidation

Leverage decides two things: how much collateral you post for a position, and where the exchange closes it for you. It does not decide how much you lose at your stop — the size does. Where it bites is liquidation: the higher the leverage, the closer the liquidation price sits to the entry, and if it sits closer than your stop, the stop never gets its turn.

For a linear USDT-margined contract in isolated margin, ignoring fees, the liquidation price can be estimated from the leverage and the exchange's maintenance margin rate (MMR) for the position's size tier:

With an MMR of 0.5% and an entry at 100,000, a 10× long liquidates at about 100,000 × 0.9 ÷ 0.995 ≈ 90,452 — roughly 9.55% below the entry. At 100× the same formula gives ≈ 99,497, about 0.50% below the entry: a move BTC often makes within an hour. A stop at 98,000 sits comfortably above the 10× liquidation and hopelessly below the 100× one.

The real liquidation price depends on the exchange's maintenance-margin tiers (the rate rises with position size), on the mark price rather than the last trade, on fees and on any margin you add later: treat the formula as an estimate and the number in your exchange's order form as the truth. The liquidation price calculator lets you try different leverage and MMR values before placing the order. The practical rule: the stop belongs well before the liquidation price, with a buffer for wicks and funding.

R-multiples: measure results in units of risk

Once every trade has a planned loss, that loss becomes a unit. One R is the amount you would lose at the stop — 100 USDT in the example above. A trade that closes at the take-profit made +3R; one that stops out made −1R; one closed early for 150 USDT made +1.5R. Results in R are comparable across accounts of different sizes and across months with different risk amounts, which results in USDT are not.

R also gives a clean definition of expectancy: win rate × average win (in R) − loss rate × average loss (in R). A system that wins 40% of the time with +2R winners and −1R losers expects 0.4 × 2 − 0.6 × 1 = +0.2R per trade. Positive expectancy with a 40% win rate is normal; what matters is the ratio, not the hit rate.

The ratio needed to break even follows from the same formula: reward-to-risk = (1 − p) ÷ p, where p is the win rate.

Any edge sits above these numbers, and it has to cover fees as well. A setup that wins half its trades needs winners larger than losers just to stand still; add roughly 0.1% of position value in fees per round trip and the bar moves up again.

Where the stop goes

There are two honest ways to place a stop. A structure-based stop goes just beyond the level that would prove the idea wrong — below the swing low a long was built on, above the range high a short was sold against. A fixed-percent stop (say 2% from the entry) is simpler and fits trades without a clear level, but it has no relation to the chart.

Whichever you use, the order of operations never changes: first the stop, then the size. If a structure-based stop is far away, the position gets smaller; if the result is too small to be worth the fees, the trade is skipped. Widening a stop to fit a size you already like inverts the whole method.

Two related habits are worth naming. Moving a stop further away after the entry turns a planned −1R into an unplanned −2R or −3R; the planned loss is the one number that should never be edited while the position is open. Moving the stop to break-even early caps the loss at zero but tends to produce more small losses on trades that would have worked — a trade-off rather than a free improvement, and whether it pays depends on the setup.

Fees and funding eat small edges

A round trip costs about twice the taker fee, because the fee is charged on the position value at entry and again at exit. With 0.05% per side that is roughly 0.1% of the position value — small on one trade, decisive for a strategy whose average winner is 0.5%.

Worked example, matching the PnL calculator: a long of 0.05 BTC from 100,000 to 102,000 at 10× leverage with a 0.05% fee per side. The gross profit is (102,000 − 100,000) × 0.05 = 100 USDT. The fees are (5,000 + 5,100) × 0.05% = 5.05 USDT, so the net is 94.95 USDT. The margin at 10× is 500 USDT, which makes the ROE 94.95 ÷ 500 = 18.99% — on a 2% move.

Perpetual contracts add funding, exchanged between longs and shorts at fixed intervals — typically every eight hours, on some contracts more often. On a trade that lasts minutes it is irrelevant; on one held for days against the crowd it can cost more than the entry and exit fees combined. Record it as part of the trade's result, not as a separate line that is easy to ignore.

Drawdowns: the maths of getting back

Losses and gains are not symmetric. A 10% loss needs an 11.1% gain to get back to the starting balance, and the deeper the hole, the steeper the climb:

This is the arithmetic behind the 1–2% rule: a run of small losses leaves a recoverable account, while a few large ones leave a hole that needs a different, better trader to climb out of. Many traders add a circuit breaker on top — a daily loss limit of, say, 3R or a weekly one of 6R — after which they stop trading until the next period. The exact numbers are personal; the point is that the limit is decided calmly in advance, not during the losing day.

Several positions at once

Risk per trade assumes trades are independent. In crypto they rarely are: most altcoins move with BTC, and a long BTC, a long ETH and a long SOL opened on the same day are closer to one position at three times the size than to three separate ideas. A single market-wide drop can hit all three stops at once.

For that reason many traders cap the total open risk — the sum of planned losses across every open position — at around 3–5% of the account, and count correlated positions as one when applying the cap. The cap also protects against the slow version of the problem: adding a fourth and fifth trade because the first three look fine.

Put it in the journal

None of this works without a record. A risk plan that lives in your head gets renegotiated every time the position is red. What goes in the journal, per trade:

Reviewed per setup, those fields answer the questions that matter: which setups have positive expectancy, whether the losses are planned −1Rs or unplanned −3Rs, and how much of a drawdown came from broken rules rather than failed ideas. If you are starting from scratch, how to keep a trading journal covers the fields and the review routine.

PnLTrack keeps this record with little typing: it syncs trades about every minute from Binance, Bybit, OKX, MEXC, BingX, Hyperliquid and Propr through read-only API keys, takes a plan with a stop and a target for each trade, shows the R-multiple of every trade that has a stop, and breaks statistics down per strategy. The FREE plan is free forever; see the supported exchanges and pricing for details.