Fees, slippage and perpetual funding compound against a strategy. See the turnover × cost math and what a 10,000 USDT BTC long paid in funding.
Risk Metrics · Updated 2026-10-08
Every trade has a price on the chart and a price you actually get. The difference is made of explicit fees and implicit costs, and on perpetual futures there is a third cost that accrues while you simply hold the position. Each looks small in isolation. Over hundreds of trades or months of holding, they compound into a meaningful share of the return.
Trading fees are charged by the venue as a percentage of the traded value, or notional. Many venues charge different rates for orders that add liquidity (maker) and orders that take it (taker). Fee schedules vary by exchange, account tier and product, and they change over time, so always check the venue's own current fee page rather than relying on a remembered number.
Slippage is the gap between the expected price and the executed price. It comes from the bid-ask spread, from order size relative to available liquidity, and from price moving while the order fills. It never appears on a fee statement but is just as real: a market order that fills 0.05% worse than the quoted price costs 0.05% of notional, exactly like a fee.
Funding applies to perpetual futures. Because a perpetual contract has no expiry, exchanges use periodic funding payments between longs and shorts to keep its price close to the spot price. When the funding rate is positive, longs pay shorts; when it is negative, shorts pay longs. Payments are exchanged at each settlement, commonly every 8 hours, and are calculated on the position's value.
Turnover is the total value traded over a period divided by capital. Selling the whole portfolio and buying a new one is one full rotation, which means two trades, each equal to 100% of capital. The approximate annual cost drag is simply turnover × cost per side. For a more exact figure, multiply the fraction that survives each trade: capital after costs = (1 − cost)^(number of full-size trades).
The key point is that cost scales with activity, not with profit. A strategy that trades twice as often pays roughly twice the drag, whether or not the extra trades add anything.
The funding rates page records BTCUSDT perpetual funding from the Binance and Bybit public APIs. As of 7 October 2026, the 30-day window on each venue contains 90 settlements at 8-hour intervals. The sum of those 90 rates was 0.411137% on Binance and 0.370401% on Bybit. Funding rates change at every settlement, so treat these as a dated snapshot, not a quote.
Funding paid = position value × sum of rates over the holding period. For a 10,000 USDT BTC long held through those 30 days at a constant notional, Binance comes to 10,000 × 0.411137% = 41.11 USDT and Bybit to 10,000 × 0.370401% = 37.04 USDT. The difference is about 4.07 USDT for the month. A short of the same size would have received those amounts instead.
Annualized figures follow the same logic. The average 8-hour rate on Binance over the window was 0.004568%, and 0.004568% × 3 settlements a day × 365 days ≈ 5.00% a year; Bybit's 30-day annualized figure was 4.51%. Relative to margin, the cost looks larger. At 5x leverage the 10,000 USDT position uses 2,000 USDT of margin, so the 41.11 USDT on Binance is about 2.06% of that margin in one month, before any price move. The direction can also flip: 90.0% of Binance's 30-day settlements were positive, but the latest one, at 00:00 UTC on 7 October 2026, was slightly negative at −0.000743%. The funding fee calculator lets you plug in any rate, interval and holding period.
The strategy backtests on this site charge a modelled cost on every trade, between 0.10% and 0.25% of traded value depending on the strategy, and one strategy instead charges 0.2% multiplied by its turnover. The headline CAGR and drawdown figures are therefore already after these costs. To see why the setting matters, consider the following illustration, which is arithmetic only and does not describe any specific strategy.
Suppose a strategy fully rotates its portfolio 24 times a year, which means 48 trades each equal to the whole capital. At 0.25% per side, the rough drag is 48 × 0.25% = 12% of capital a year. Compounded exactly, capital keeps (1 − 0.0025)^48 = 0.8868 of its value, an annual drag of about 11.3%. At 0.10% per side, the same activity gives (1 − 0.001)^48 = 0.9531, a drag of about 4.7%. Over ten years, the 0.25% case leaves 0.8868^10 ≈ 0.30, roughly 30% of the ending value the same strategy would have reached at zero cost, while the 0.10% case leaves about 62%. Small per-trade numbers become large when multiplied by frequency and time.
The same arithmetic applies to Live Paper Trading, where paper fills charge 0.04% of notional per trade and pairs-spread templates pay it on both legs. A grid template that fills hundreds of times can see that small rate add up to a visible share of its result.
Judging an active strategy on gross returns. If an average round trip earns 0.15% before costs and costs 0.10% per side, it keeps 0.15% − 0.20% = −0.05%. The strategy is a loser that looks like a winner until costs are added.
Ignoring slippage because it does not appear on the statement. Models that assume fills exactly at the signal or closing price understate costs for larger orders and thin markets.
Treating funding as fixed. A 30-day average is history. Funding can turn strongly positive in crowded rallies and negative in sell-offs, and a position held for months pays or receives every one of those settlements.
Forgetting that leverage magnifies funding relative to your own capital. Funding is charged on the position value, not on the margin you posted.
Costs are certain; returns are not. Before trusting any return curve, ask what fees, slippage and funding it assumes and how often it trades. Check each exchange's current fee schedule and product availability in your country yourself. Backtests and paper results are hypothetical and may understate real execution costs. This is educational material, not investment advice.
Educational content, not investment advice. Guide content and league rankings are independent of any exchange or sponsor relationship.