Position sizing for crypto traders: a two-layer risk system that survives volatility

Position sizing for crypto traders means turning a fixed risk budget (often 1–2% of account equity) and a predefined stop into an exact number of coins or contracts. The durable approach is two-layered: size every trade off stop distance, then cap portfolio heat so multiple positions cannot combine into one outsized drawdown when correlation spikes.

Key Takeaways

• A usable position size starts with a stop loss order, because the core calculation is dollars-at-risk divided by stop distance.
• The 1 percent rule is a risk budget, not a sizing method by itself. Stop distance converts that budget into units.
• ATR is a volatility equalizer: higher ATR forces wider stops and therefore smaller size, which is the point when ranges expand.
• Portfolio heat is the second layer. A 6% total open-risk cap is a common heuristic to stop “five 1% trades” becoming one correlated hit.

Position sizing and why it matters

Position sizing is the part of crypto trading risk management that standardizes worst-case loss across wildly different volatility regimes. The screen-level decision is simple: before clicking buy or sell, the trader decides where the trade is proven wrong and how much account equity can be lost if price hits that level. Everything else is arithmetic.

Crypto makes this non-negotiable because volatility regimes change fast. The Medium Kelly piece cites academic work (Baur, Hong, Lee 2018) that found Bitcoin volatility is substantially higher than major fiat currencies and gold, which is another way of saying the same nominal position can behave like three different risks depending on the week. A fixed “$5,000 BTC position” is not a fixed risk. A fixed loss-at-stop is.

Most guides stall at “risk 1% and set a stop.” That is only layer one. Layer two is the book-level constraint: when majors and alts flush together, correlation goes to 1 and a cluster of individually “small” risks can land on the account at once. Chart Guys explicitly pairs per-trade sizing with portfolio heat and references a 6% aggregate risk rule.

This is also where common tools get reframed. A crypto position size calculator is not a magic box. It is just automating three inputs the trader must own: account risk budget, entry, and stop distance. If any of those are fuzzy, the output is theater.

Core math using fixed-percent risk

The sizing math starts with a constraint, not a forecast. Chart Guys frames the core calculation as: position size equals maximum dollar risk per trade divided by per-unit risk, where per-unit risk is defined by stop distance (for a long, entry price minus stop price). That is the whole engine.

The 1% and 2% rules are common defaults for setting that maximum dollar risk. Chart Guys and CryptoTailor both reference risking 1–2% of total account value per trade. The important nuance is what that percent applies to. It is not “1% of the position.” It is “1% of the account if the stop is hit.”

A concrete workflow looks like this:

1. Set the risk budget. Example: a $25,000 account using the 1 percent rule means $250 max loss on the trade. 2. Define the invalidation level. The stop loss order price is the line where the idea is wrong, not a number chosen to fit a desired size. 3. Measure stop distance. If entry is $100 and the stop is $95, the per-unit risk is $5. 4. Compute units. $250 ÷ $5 = 50 units.

That “50 units” is the position size (position-size) that makes the loss at the stop equal to the risk budget. If the stop has to be wider because the market is swinging, the units must shrink. That is the point of the system.

This is also where leverage (leverage) quietly breaks people. Leverage changes liquidation mechanics and margin usage, but it does not change the arithmetic of dollars-at-risk at the stop. If the stop is far and the size is large, the account is still exposed, even if the margin posted looks small.

Volatility-based stops with ATR

ATR enters the process as a stop-distance input, not as a prediction tool. MOSS makes the key comparison explicit: a coin with daily ATR around 8% implies wider stops and therefore smaller position sizes than a coin with ATR around 2%. That is exactly what a sizing system should do when a market’s daily range expands.

The mistake is treating ATR as a “better stop” that magically improves win rate. ATR is a volatility ruler. It tells the trader how noisy the tape has been, so the stop can be placed far enough away to avoid getting clipped by routine movement. Once the stop distance expands, the fixed-percent sizing math forces the position size down.

CryptoTailor describes ATR-based position sizing approaches that risk a small percent of capital (often 1–2%) and adjust size dynamically with ATR. The implementation varies by trader, but the screen-level behavior is consistent: when ATR rises, the stop distance used in the sizing calculation rises, and the number of units falls.

Historical volatility is a parallel input when ATR is not the chosen tool. MOSS references using the standard deviation of returns and suggests 30–90 day windows for measuring it. The practical use is the same as ATR: it is a regime filter. A coin that has been whipping around for the last 30–90 days should not be sized the same as a coin that has been quiet, because the stop distance required to survive normal noise is different.

For traders running multiple markets, this is where sizing stops being cosmetic. Volatility-adjusted stops prevent the common failure mode of “same notional across alts,” where the highest-vol names dominate the PnL distribution and turn the book into a disguised single bet.

Edge-based sizing with fractional Kelly

Kelly sizing is a different animal because it sizes from estimated edge rather than from a fixed risk budget. The Medium article traces Kelly back to John L. Kelly Jr.’s 1956 framework and uses the standard inputs: probability of winning and payoff odds. The output is a fraction of capital to allocate.

The crypto-specific problem is that Kelly can recommend extreme allocations when the assumed edge looks strong. The Medium piece provides worked examples where full Kelly produces recommendations above 100% of capital, which implies leverage or margin. That is not a rounding error. It is the formula telling the trader to press hard.

Fractional Kelly is the compromise that shows up on professional desks because the inputs are fragile. The Medium article states that many professionals use 25–50% of the Kelly recommendation to reduce risk while giving up only modest returns, citing research. Quarter-Kelly and half-Kelly are common labels for that.

This is where the two-layer framework matters. Fixed-percent risk sizing standardizes loss at the stop. Fractional Kelly is only sensible after a trader can defend an edge estimate and still respects book-level constraints like portfolio heat. If the win-rate and payoff assumptions are wrong, full Kelly does not fail gracefully. Crypto’s volatility regime shifts make yesterday’s probabilities unreliable, which is why the Medium piece emphasizes the difficulty of probability estimation and the behavioral tendency toward overconfidence.

Kelly also interacts with r multiple (r-multiple) and risk reward ratio (risk-reward-ratio). Those two metrics are how many traders summarize payoff distribution after the fact. Kelly needs a forward-looking estimate of that distribution, which is a higher bar than logging a few good trades.

Portfolio heat and common sizing mistakes

Portfolio heat is the number that stops “disciplined” single-trade sizing from turning into undisciplined book risk. Chart Guys calls out portfolio heat as a core concept alongside per-trade sizing and references a 6% rule for aggregate exposure. Read that as: if the sum of all open trade risks at their stops is near 6% of equity, the next setup is not free.

This matters most in crypto because correlation clusters. A trader can have five separate charts, five separate theses, and still be long the same factor. When the market sells off, those stops can get tagged in the same candle.

The common mistakes are mostly conceptual, not mathematical:

1. Treating “1% risk” as the position size. The percent is only the risk budget. Stop distance determines units via the dollar risk ÷ stop distance formula (Chart Guys). 2. Placing the stop to fit a desired size. If the stop has to be wider because ATR is elevated, the system demands a smaller position size. Keeping size and moving the stop is how drawdowns accelerate. 3. Ignoring portfolio heat. “I only risk 1% per trade” is not a defense if six positions are open and the book is effectively one correlated bet. The 6% aggregate heuristic exists to prevent that (Chart Guys). 4. Letting Kelly outputs override risk controls. Full Kelly can exceed 100% allocation in the Medium examples, which is a direct path to leverage exposure when the edge estimate is wrong.

A clean operational definition helps: if the stop loss order is not written down, there is no position size. There is only a notional.

A practical sizing checklist

This checklist is the repeatable pre-trade sequence that keeps sizing consistent across coins, volatility regimes, and multiple open positions.

1. Write the trade invalidation level and the stop loss order price. If the stop cannot be justified on the chart, the trade is not ready to size. 2. Pick the per-trade risk budget. Use a fixed-percent rule like 1–2% of account equity as the default risk per trade (Chart Guys. CryptoTailor). 3. Convert the risk budget into dollars. Account equity × risk percent equals maximum loss if stopped. 4. Measure stop distance from entry to stop. That distance is the per-unit risk used in the sizing formula (Chart Guys). 5. Compute units with a crypto position size calculator or a spreadsheet. The calculator is only automating dollars-at-risk ÷ stop distance. 6. Sanity-check volatility. If ATR or recent volatility is elevated, expect the stop distance to be wider and the position size to be smaller (MOSS. CryptoTailor). Do not “fix” that by tightening the stop to keep size. 7. Check portfolio heat before placing the order. Sum the open risks at stops across positions and compare to a cap such as 6% (Chart Guys). 8. Record the expected r multiple and risk reward ratio for the setup. This is the data that later tells whether any edge-based sizing like fractional Kelly is even defensible.

This is the point where crypto trading risk management stops being a slogan and becomes a daily habit. The trader is not trying to maximize returns on the next trade. The trader is trying to standardize worst-case loss so the account survives long enough for any edge to show up.

The Take

I’ve watched traders do immaculate 1% sizing on each chart and still eat a 5–6% hit in one ugly flush because the book was five versions of the same bet. The tell is always the same: they can quote the 1 percent rule, but they can’t tell you their portfolio heat before they add the next position.

The habit that keeps this clean is boring. I size only after the stop loss order is written, then I treat the portfolio heat cap as a circuit breaker. If ATR is screaming and the math spits out a smaller position size, that is the market charging more rent for risk. Pay it by trading smaller, not by pretending volatility went away.

Sources

• The Chart Guys
• MOSS
• CryptoTailor
• Medium