The real cost of crypto fees and slippage

“Binance charges 0.1% per trade” is true and misleading. The actual cost of executing a strategy is 2 to 4 times higher. If your backtest uses 0.1% or 0.2% round-trip, it's under-costing reality by enough to turn a profitable strategy into a losing one.

This is every component of real per-trade cost, with actual numbers from our own calibration work.

Component 1: exchange fees

This is the advertised number. Binance spot taker is 0.1%. Binance perp taker is 0.04%. Kraken taker is 0.26%. Coinbase Advanced taker is 0.6%. Maker fees (for limit orders that don't cross the spread) are lower, often 20-50% of taker.

Round-trip, meaning one entry + one exit, you pay this twice. Binance spot taker round-trip is 0.2%. Binance perp taker is 0.08%. The perp number is lower partly because perps have separate funding-rate cash flow that can offset or add to cost depending on direction.

If your backtest uses Binance spot rates, round-trip fee assumption should be at least 0.2% for taker, 0.15% for realistic mixed maker/taker fills.

Component 2: bid-ask spread

You buy at the ask. You sell at the bid. The difference is the spread, and it is an invisible cost that nobody charges you explicitly. On major coins at Binance it's often 1-5 basis points (0.01-0.05%). On smaller tier-2 or tier-3 alts, it can be 10-50 bps (0.1-0.5%).

Round-trip spread cost: cross the spread once on entry, once on exit. If BTC spread is 2 bps one-way, that's 0.04% round-trip. If your strategy trades ARB or SUI with 10 bps spreads, that's 0.2% round-trip. For long-tail alts, 0.4% is not unheard of.

Spreads scale with liquidity tier. Top-2 majors (BTC, ETH) are tight enough to be nearly negligible. Mid-cap majors are manageable. Small alts can double or triple the round-trip cost by themselves. Our backtester uses per-coin median spreads sampled from live L2 books so the right tier cost is baked into every test automatically.

Component 3: execution slippage

Even after paying the fee and crossing the spread, your fill price is often worse than the best bid or ask you saw. This is because between sending the order and it hitting the book, other orders land first, and because market orders walk the book (consume multiple price levels).

We calibrate this against our own live trades continuously. The rough finding: real execution slippage runs noticeably higher than flat estimates suggest. For a typical retail position size on BTC, slippage is in the single-digit bps per side during calm markets. On stressed moves or tier-2 coins, it can blow out by a factor of several.

Round-trip slippage cost: also doubled (once entering, once exiting). So 0.1-0.4% for typical retail flows on majors, more for alts or fast markets.

Component 4: funding rate (perps only)

If you're trading perpetual futures, funding is paid every 8 hours. Binance BTCUSDT perp funding averages around 0.01% per 8h. That's ~0.03% per day or ~11% per year of being long. Funding can spike to 0.1% per 8h during euphoric bull periods (so 10× average).

Directional strategies that hold perps for multiple days need to bake in funding. A long position held for 3 days owes roughly 0.09% in funding at average rates, assuming positive funding (shorts paying longs). If funding is negative, longs get paid.

The simplification: “spot” strategies don't pay funding. Perp strategies should assume 0.03% per day of holding as a headwind on the long side.

Component 5: stop overshoot

This one bites quietly. Your stop loss is set at $49,000. The price prints $49,100, $49,000, $48,900, $49,050. Your stop fires somewhere between $48,900 and $49,050. Not at $49,000 exactly.

On average, we calibrated stop fills at about 0.15% worse than the stop level. Stops are painful precisely when you need them most, which is when the market is moving fast and the nearest resting liquidity is several ticks below your trigger.

If your strategy hits stop on 40% of trades and stop overshoot is 0.15%, you lose an extra 0.06% per average trade just from this. Over a year of 200 trades, that's -12% of notional.

Putting it together

For a typical retail strategy trading BTC perps on Binance with medium turnover:

Add it all up. Exchange fee, spread, execution slippage, funding on perp holds, occasional stop overshoot. For a typical retail-size strategy on majors, the realistic round-trip cost lands meaningfully higher than the advertised exchange fee suggests. On alts or during stress, meaningfully higher again. On high-turnover strategies the whole year's edge can disappear into costs.

The exact multiplier depends on your coin, your position size, your typical holding time, how often your strategy hits its stop, and whether you're on spot or perps. There isn't one number that applies to everything — which is precisely why a realistic backtester has to model each component separately per trade rather than assume a flat commission.

This is the entire reason a pretty backtest and a losing live strategy can look identical at first glance. The difference is all friction.

How to measure it yourself

The rigorous way is to trade small and log every fill. For each trade, record:

• Signal price (the price when your strategy said “enter”)
• Fill price (what you actually got)
• Theoretical exit (where the strategy's backtest said the exit would hit)
• Actual exit fill

After 100 trades, compute the gap. That's your real friction number, measured from your own order flow on your own account at your own size.

The shortcut is to use a tool that's already calibrated this. Our backtester uses the 2× slippage multiplier we empirically derived, per-coin spreads from live L2 samples, and a 0.15% stop overshoot. We publish the calibration updates when they change.

The actionable takeaway

If your backtest assumes less than 0.25% round-trip cost on majors or 0.4% on alts, your backtest is lying to you about a strategy's live performance. Any strategy whose edge evaporates when you bump the fee assumption to realistic values never had a real edge. It was earning the friction you hadn't accounted for.

The rule of thumb: if your strategy doesn't work at 0.4% round-trip, it doesn't work.