Position Sizing and Leverage: 3x Turned +221% Into -58%
Written by chwoo · Crypto trader, 7 years. Running an automated futures bot since spring 2026. · · updated
Ask a new trader why they lost money and they will talk about their entry. Ask a veteran and they will talk about size. Position sizing, meaning how much of your account you risk on each trade, quietly decides whether a real edge compounds into wealth or a small losing streak wipes you out. It matters more than the entry, and it is the part most people never test.
Rather than repeat that claim, this guide measures it with the site's engine. One of the results came out the opposite of what we expected, and that is the most useful part of the page.
An edge and survival are different problems
A strategy can be profitable on average and still ruin you if one position is too large when the inevitable losing streak arrives. A backtest reports the average outcome; you live the path. Two traders running the identical strategy can end the year one rich and one liquidated. The difference is how much they bet per trade.
Leverage on a losing strategy
We started with a strategy that already loses: long when RSI(14) crosses above 30, 3% take profit and 2% stop loss, BTC on the 1-hour timeframe, 0.06% fee and 0.02% slippage per side, 90% position size, starting from $10,000. Only the leverage changed.
| Leverage | ROI | MDD | Final equity | Trades |
|---|---|---|---|---|
| 1x | -79.8% | 85.2% | $2,016 | 696 |
| 2x | -97.1% | 98.4% | $291 | 696 |
| 3x | -99.7% | 99.9% | $30 | 696 |
| 5x | -100.0% | 100.0% | $0.12 | 696 |
| 10x | -100.0% | 100.0% | $0.00 | 696 |
Open the strategy this guide measures in the backtester →
Notice that all five rows have the same 696 trades. Leverage changes nothing about when you buy or sell. It takes the same trades in the same order and multiplies the size of each outcome. The difference between $2,016 left and nothing left is that multiplication alone.
Leverage on a winning strategy
This is the part that matters. We took the one preset in this site's gallery that performs strongly, Supertrend momentum trading both directions, on BTC 4-hour candles. This is a strategy that actually makes money.
| Leverage | ROI | MDD | Final equity | Trades |
|---|---|---|---|---|
| 1x | 221.3% | 42.6% | $32,132 | 376 |
| 2x | 111.6% | 68.2% | $21,163 | 376 |
| 3x | -58.5% | 93.6% | $4,152 | 376 |
| 5x | -99.9% | 100.0% | $8 | 376 |
| 10x | -100.0% | 100.0% | $0.00 | 376 |
Doubling the leverage should double the return, and instead it halved it, from +221.3% to +111.6%. At 3x the sign flips outright to -58.5%. Same strategy, same 376 trades, and it goes from winning to losing.
The cause is the asymmetry of compounding. Lose 50% and then gain 50% and you are not back to even, you are at 75%. Leverage enlarges the percentage decline through every losing stretch, and you enter the next trade with the smaller capital that decline left behind, so the gain needed to recover grows. That deficit accumulates while the drawdown widens from 42.6% to 93.6%. A backtest that headlines only the return hides this process entirely.
The surprise: nothing was ever liquidated
Across this entire sweep, zero positions were liquidated. The engine models isolated-margin liquidation, booking a loss larger than the margin as a total loss of that margin, and not one trade reached it. The accounts at 5x and 10x still went to essentially nothing.
Here is why that is possible. The stop is 2% and liquidation requires losing the entire margin, so an individual trade blowing up spectacularly is rare. The account does not die in one hit, it gets ground down. Several hundred small losses, each multiplied by leverage and compounded, erase the capital without a single liquidation notice.
That directly contradicts the most common belief about leverage. People think the danger is the liquidation price and that avoiding it keeps you safe. What actually empties the account is not the liquidation line, it is that every ordinary stop-out now costs three or five times what it used to.
Moving size instead of leverage
We took the same winning strategy, fixed leverage at 1x, and changed only the position size.
| Position size | ROI | MDD | Final equity |
|---|---|---|---|
| 10% | 25.8% | 5.9% | $12,575 |
| 25% | 68.3% | 14.0% | $16,829 |
| 50% | 141.0% | 26.3% | $24,103 |
| 90% | 221.3% | 42.6% | $32,132 |
| 100% | 228.6% | 46.2% | $32,856 |
Size behaves differently from leverage. Growing it fivefold from 10% to 100% never flips the sign; return and drawdown rise alongside each other. And it loses efficiency at the top: going from 90% to 100% adds 7.3 percentage points of return and 3.6 points of drawdown. Meanwhile 50% delivers 141% at a 26.3% drawdown, a better ratio of return to pain than 90% gives.
This table does not tell you which row is correct. If you can sit through a 42.6% drawdown, 90% is defensible; if your hands shake at 26%, 50% is. What a backtest provides is the exchange rate. The choice is yours.
3x does not care which coin you picked
We ran the same strategy at 3x and 90% size across six coins. At 1x this strategy is profitable on five of the six.
| Coin | ROI | MDD | Trades |
|---|---|---|---|
| BTC | -58.5% | 93.6% | 376 |
| ETH | -69.0% | 99.8% | 380 |
| SOL | -97.6% | 99.2% | 304 |
| XRP | -100.0% | 100.0% | 352 |
| DOGE | -78.9% | 99.8% | 290 |
| ADA | -99.8% | 100.0% | 378 |
Three times leverage turned this strategy's record across coins from five wins and one loss into zero and six. The more volatile the coin, the harder it lands: XRP and ADA reach a 100% drawdown, meaning no account left.
The three levers
- Position size: the share of capital committed to one trade. As the table shows, it moves return and drawdown together and gently.
- Leverage: a multiplier on exposure. It multiplies volatility and shaves return through compounding. It is the only lever here that flips the sign.
- Risk per trade: what you lose when the stop is hit. Professionals often cap this at a small percentage of the account, frequently quoted around 1 to 2%, so no single trade can be fatal.
Testing it here yourself
The capital panel sets initial capital, fee and slippage per side, leverage and position size. Run your strategy at 1x, then 2x, then higher, and watch what happens to maximum drawdown and final equity. Read only the return and you will walk straight into the trap in the tables above.
Then read the result next to this site's Kelly and compounding calculators. Kelly gives a theoretical ceiling for sizing from your win rate and payoff ratio; full Kelly is brutally volatile, so most practitioners use a fraction of it. The compounding maths shows that surviving drawdowns, rather than maximising any single year, is what actually grows an account.