Do Fees and Slippage Really Matter? We Ran 8 Strategies Twice
Written by chwoo · Crypto trader, 7 years. Running an automated futures bot since spring 2026. · · updated
Plenty of strategies are profitable in a world without costs, and that world does not exist. Fees, slippage and leverage are not footnotes; for active strategies they routinely decide the sign of the final return. If your backtest ignores them, it is measuring a fantasy.
Rather than repeat that, we ran all eight of this site's presets twice, once at zero cost and once at realistic cost, and counted how many flipped. The answer was half of them.
Strategies that only work at zero cost really exist
Each of the eight gallery presets was run on BTC at its own default timeframe, twice: once with fees and slippage set to zero, once with realistic values (0.06% fee plus 0.02% slippage per side). 1x leverage, 90% position size, data through 31 August 2026.
| Strategy | Trades | Zero cost | Real cost | Gap |
|---|---|---|---|---|
| RSI with trend filter | 80 | +16.5% | -0.2% | 16.7pp |
| Donchian breakout | 477 | +16.6% | -41.5% | 58.1pp |
| Golden cross | 40 | +207.9% | +190.7% | 17.2pp |
| Supertrend momentum | 376 | +443.1% | +221.3% | 221.7pp |
| RSI in at 30, out at 70 | 420 | -24.0% | -59.1% | 35.0pp |
| RSI oversold bounce | 676 | -58.3% | -83.2% | 24.8pp |
| MACD cross | 1,743 | -81.0% | -98.6% | 17.6pp |
| Bollinger mean reversion | 2,126 | -61.6% | -98.1% | 36.5pp |
Open the strategy this guide measures in the backtester →
The trend-filtered RSI goes from +16.5% to -0.2%, and the Donchian breakout from +16.6% to -41.5%. Not one rule changed. Fees and slippage were simply switched on. If you had discovered either of those strategies with a cost-free backtest, you would have discovered an edge that does not exist.
The two survivors paid as well. Supertrend momentum falls from +443.1% to +221.3%, giving up half its return to costs. That is the strongest preset on this site, and costs still take half.
Trade frequency multiplies the cost
How much costs take is not set by the rate alone; it is set together with how many times you pay it. We ran an identical rule (RSI crossing above 30, 3% take profit and 2% stop) changing nothing but the timeframe.
| Timeframe | Trades | Zero cost | Real cost | Gap |
|---|---|---|---|---|
| 15 minutes | 1,615 | -5.2% | -91.7% | 86.6pp |
| 1 hour | 696 | -44.0% | -79.8% | 35.9pp |
| 4 hours | 223 | -23.6% | -43.3% | 19.8pp |
| 1 day | 28 | -10.0% | -13.1% | 3.1pp |
Look at the 15-minute row again. At zero cost it is -5.2%, very nearly breakeven. At realistic cost it is -91.7%. Most of that strategy's loss is not a judgement about the market, it is the result of paying 0.16% round trip 1,615 times.
At the other end, the daily timeframe trades 28 times so costs take only 3.1 percentage points. Before concluding "use the daily then", remember that 28 trades is far too few to be a statistic. Reducing costs by burning the sample is not the same thing as beating them.
How sensitive is the fee rate
On the 696-trade strategy, holding slippage at 0.02% and moving only the fee.
| Fee per side | ROI | Final equity |
|---|---|---|
| 0% | -57.2% | $4,278 |
| 0.02% | -66.7% | $3,329 |
| 0.04% | -74.1% | $2,591 |
| 0.06% | -79.8% | $2,016 |
| 0.10% | -87.8% | $1,220 |
The 0.02 percentage points between maker and taker, or between exchange tiers, look trivial until they are multiplied by 696 trades, at which point the final equity is $3,329 or $2,016. That is why this site defaults to 0.06% per side: a conservative futures taker rate, adjustable but not to be ignored.
Slippage is quieter and hurts more
Now holding the fee at 0.06% and moving only slippage.
| Slippage per side | ROI | Final equity |
|---|---|---|
| 0% | -73.5% | $2,648 |
| 0.02% | -79.8% | $2,016 |
| 0.05% | -84.6% | $1,544 |
| 0.10% | -92.6% | $744 |
| 0.20% | -98.3% | $166 |
A fee appears on your statement; slippage does not. That is why most backtests leave it at zero. The table shows how large an assumption that zero is. On altcoins or in volatile moments, 0.2% per side is realistic, and at that level this strategy's $10,000 ends as $166.
The more a strategy trades breakouts or panic candles, the more slippage it pays, because the candle that generated the signal is usually the candle with the thinnest order book. This site's 0.02% default assumes market orders on majors; if you are testing a breakout system, the honest move is to raise it and run again.
Leverage is not a cost, it is an amplifier
Fees and slippage are amounts that leave the account on every trade. Leverage is different in kind: it multiplies the outcome of the same trades. So it does not solve a cost problem, it enlarges one, because fees scale with exposure too.
How leverage actually ends an account is measured in a separate guide. In short, even the strongest strategy on this site turns +221% into -58.5% at 3x, and not one liquidation notice appears along the way. The account does not die in one hit; it gets ground down.
A practical checklist
- Only judge with costs switched on. Two of the eight above flipped sign. Looking good at zero cost is not information.
- Read the trade count as a cost multiplier. A 0.16% round trip is 16% over 100 trades and 160% over 1,000.
- Raise slippage and run it again. If something works at 0.02% and collapses at 0.05%, its edge is resting on fill quality.
- Check your exchange's rate. 0.02 of a point per side changed $1,300 across 696 trades.
- Do not try to out-earn costs with leverage. Larger exposure means larger fees too.
References
- Slippage (finance) — Why a market order fills away from the price you saw.
- Risk of ruin — The leverage problem stated formally.
- Binance public market data (data.binance.vision) — The candle archive every number on this site is computed from.