Crypto perpetual funding is a periodic cash flow attached to an open contract, separate from entry and exit commission. Its amount depends on the applicable funding rate, position value, direction and settlement rules. A small percentage can become material relative to posted collateral, and a rate visible now does not promise the same payment throughout a planned holding period.
Before calling a backtest profitable or a position cheap to hold, add funding as its own ledger. This article was prepared with AI assistance and uses an original hypothetical USDT-linear example, not observed rates, a historical performance study or an investment recommendation.
First identify what the percentage applies to
A funding rate generally describes a payment calculation for a particular contract and interval, not a universal annual yield. The settlement asset and definition of position value matter.
Bybit documents funding payments between long and short perpetual holders at settlement. Its USDT-contract example calculates funding from quantity multiplied by mark price and the applicable rate; positive rates make longs pay shorts, with the direction reversed for negative rates. Inverse contracts use a different position-value formula and settlement unit. Bybit: funding-fee calculation.
That scope is deliberate. Multiplying every contract count by a dollar price can be wrong when a contract is inverse or uses another multiplier. Determine what one unit means before computing notional.
Also distinguish position value from collateral. A 10,000 USDT notional is not the same base as a hypothetical 1,000 USDT collateral allocation. A funding charge calculated on the former will be a larger percentage of the latter.
This is one reason reducing a position and reducing a collateral allocation are different actions. See the reduce-only order explanation before treating an account balance change as a position close.
Three settlements, three different amounts
Construct a long position equivalent to five ETH in a USDT-linear perpetual. Assume it remains five ETH throughout three funding settlements. At each settlement, assume the position is definitely eligible under the venue's rules. Ignore all other account activity initially.
The following marks and rates are invented for the worksheet. A positive cash cost means the long pays; a negative cost means it receives.
| Settlement observation | Mark, USDT/ETH | Position value | Funding rate | Long's cash cost |
|---|---|---|---|---|
| A | 2,000 | 10,000 USDT | +0.010% | +1.0000 USDT |
| B | 1,950 | 9,750 USDT | +0.030% | +2.9250 USDT |
| C | 2,050 | 10,250 USDT | −0.005% | −0.5125 USDT |
For B, the decimal rate is 0.030 ÷ 100 = 0.00030. Its payment is 5 × 1,950 × 0.00030 = 2.925 USDT. Forgetting the division by 100 would overstate that payment by a factor of 100.
The net funding paid is 1.0000 + 2.9250 − 0.5125 = 3.4125 USDT. The negative third rate returns part of the earlier costs; it does not erase the first two settlements from history.
The position quantity never changed. The funding amount did because both the mark and rate changed. Holding an entry-time notional constant would produce a different result. For this constructed contract, the relevant mark is the one used at each settlement, not an arbitrary latest chart price.
If the hypothetical collateral allocation is 1,000 USDT, the net payment is 3.4125 ÷ 1,000 × 100 = 0.34125% of that allocation. That percentage is descriptive arithmetic, not the actual venue's return-on-equity metric, margin requirement or liquidation formula.
Add commission without counting it as funding
Continue the same constructed position. Suppose it entered at 2,000 and later fully exited at 2,050, with no slippage. The gross price result is 5 × (2,050 − 2,000) = 250 USDT.
Assume an invented 0.05% commission on both entry and exit, paid in USDT. Entry commission is 10,000 × 0.0005 = 5 USDT; exit commission is 10,250 × 0.0005 = 5.125 USDT. Total commission is 10.125 USDT.
The price result after only these modeled commissions and funding is 250 − 10.125 − 3.4125 = 236.4625 USDT. This calculation is not a forecast. It omits other costs and risks and assumes exact executions at the stated prices.
Preserve the components rather than displaying only the final number. A strategy with a positive price result and negative funding has a different explanation from one with poor fills and favorable funding. Combining every debit into fees can conceal the assumption that needs correction.
The slippage and depth worksheet adds another cost dimension. Real entry and exit prices already include the actual price path of those fills; do not subtract a duplicate slippage estimate from a result based on those same fills without a clearly separate benchmark purpose.
Funding intervals are instrument data
Do not assume every perpetual settles every eight hours. Bybit documents instrument-level funding intervals and changing settlement arrangements. Its instruments endpoint exposes a funding interval for the applicable products. Bybit API: instrument information.
Binance USDⓈ-M Futures separately provides funding-rate information for symbols with adjusted intervals or bounds. An empty or partial response from an adjustment endpoint should not be interpreted without reading that endpoint's scope. Binance USDⓈ-M Futures: market data.
Store the instrument, interval and next relevant settlement information with the observation time. A spreadsheet copied last month can become misleading even when its multiplication is correct. If the product changes its interval, the number of potential settlements in the holding period changes too.
For an illustrative annualization, assume a constant 0.01% every eight hours for a full 365 days. A simple noncompounded calculation gives 0.01% × 3 × 365 = 10.95%. That is a conditional mathematical scenario, not an expected annual funding return. It assumes away changing rates, intervals, position value and eligibility.
Annualizing a single elevated observation and presenting it as dependable income would add a prediction the source does not support. The same issue applies when a cost estimate assumes an unusually favorable negative rate persists for an entire trade.
Do not trade a settlement boundary as if it were a precise switch
The exact inclusion of an order placed near settlement depends on exchange processing and the product's rules. A local click time is not the same as an authoritative executed position state.
Bybit cautions that opening or closing within a short window around funding settlement does not guarantee inclusion or exclusion from that cycle. An operator should reconcile the actual settlement record rather than assuming a countdown proves a charge was avoided. Bybit: settlement timing and deductions.
This article does not recommend rapidly entering or exiting to capture or avoid funding. Such actions introduce execution costs, price movement and incomplete-order risks. A possible saved payment is not a complete comparison.
If a closing instruction is only partially executed, some position may remain eligible. A local paused label or cancel acknowledgement does not establish that all exposure disappeared. Confirm the exchange position at the relevant time and the actual cash flow afterward.
Public rate history and account cash flows answer different questions
A public funding-rate series says what rates were recorded for an instrument. It does not establish that a particular account held a particular size at each settlement. To reconstruct an account's funding costs, combine actual account records with position history and product scope.
Bybit's historical endpoint returns funding rates with settlement timestamps for supported linear and inverse contracts. Those fields are useful inputs to a reproducible model. They are not a replacement for the account's actual credited or debited amounts. Bybit API: funding-rate history.
For a backtest, specify whether funding uses recorded settlements, an assumed rate or no funding at all. Record timezone, contract, position sizes at settlement and any missing interval. An absent history should be reported as a limitation rather than silently treated as zero.
The backtest evidence guide explains why a smooth profit curve cannot validate its execution assumptions. Funding belongs in that same evidence review: a useful final number should be traceable to its inputs and timing.
Carry cost is also an account-state issue
A funding payment can change resources available to the account. Its effect depends on margin mode, collateral and other positions; do not apply the worksheet's 1,000-USDT allocation as if it were a complete margin engine.
Monitor the account's authoritative margin state and the product's rules alongside expected carry costs. The mark-price and liquidation guide separates the price used for risk calculations from the price available for an actual exit.
A plan to hold for several intervals should therefore include adverse funding assumptions and an operational review condition. That condition is a reason to reassess the position, not a promise that a close can occur at a particular price or before another settlement.
Keep actual payments separate from estimates, changing rates separate from fixed scenarios, and contract units separate from collateral. KLINEO's trading risk disclosure covers the wider losses and service limitations that this arithmetic cannot remove. A small displayed funding percentage becomes useful only when its cash amount, timing and account consequences are understood.