Crypto perpetual basis is a price gap between a perpetual contract and a stated spot reference; funding is a periodic payment rate for holding that contract. The gap describes a price relationship now, while funding reflects the venue’s calculation and settlement rules. A positive basis does not guarantee positive funding at the next settlement, and neither number guarantees a profitable trade.
This article was prepared with AI assistance. Its examples are hypothetical, use a USDT-linear contract and explain measurement rather than recommend a trading strategy.
Put a price label on the basis
For this worksheet, define basis in basis points as:
Basis = (perpetual price − spot reference) ÷ spot reference × 10,000
One basis point is 0.01%. A positive result means the chosen perpetual price is above the reference; a negative result means it is below. Record both price definitions. “BTC basis” alone leaves too much unspecified.
| Input | What to record | Why it matters |
|---|---|---|
| Perpetual price | Last trade, midpoint or executable bid/ask | A last trade can differ from the price available now |
| Spot reference | Venue index, oracle or named spot market | A composite index is not an executable spot order |
| Time | Timestamp for both observations | A moving market can create a false gap from mismatched observations |
| Units | Contract, quote currency and multiplier | Similar symbols can represent different products |
Suppose an index is 50,000 USDT and the perpetual last trade is 50,100. Last-trade basis is 100 ÷ 50,000 × 10,000 = 20 basis points. If the current perpetual midpoint is 50,080 instead, midpoint basis is 16 basis points. Both calculations can be correct because they use different prices.
A mark is another distinct reference. It may be constructed for margin and valuation purposes; it is not an offer to fill your order. The mark-price guide explains why that distinction matters to an open position.
Funding has a calculation window and a settlement clock
Funding creates a payment between holders of opposing perpetual positions. In the usual sign convention, a positive rate makes longs pay shorts; a negative rate reverses the direction. The instrument’s rules determine the relevant value, interval and eligibility.
Unlike a dated future, a perpetual has no scheduled expiry that forces its price to converge on that date. Funding supplies an economic incentive for price alignment instead. Binance’s funding explanation describes this distinction and the transfer between opposing positions. An incentive does not promise an immediate or complete closing of the gap.
Keep three observations separate: the last settled rate, the current estimate for the next settlement, and any longer-horizon scenario. A historical receipt is an account event. A displayed next rate can still change. Multiplying that estimate across a month adds an assumption about future rates.
For cash-flow calculations across several settlements, use the separate perpetual funding-cost ledger. Here the question is why a price gap and a payment rate can tell different stories.
Why a discount can coexist with positive funding
A simple basis calculation uses one price observation. A funding calculation may use an average premium index, an interest component, a dampening band and caps. Its premium index may also use impact bid and ask prices for a specified notional, rather than the last trade.
For example, Bybit’s funding methodology describes a weighted premium history, interest component, dampening band and contract limits. Under its illustrated eight-hour setup, an average premium of −0.02% and interest of +0.01% produce a +0.03% adjustment inside the ±0.05% band. The resulting funding rate is +0.01%, assuming caps do not bind. A negative average premium therefore need not produce negative funding.
That formula example is separate from the following invented screen observations. These observations do not reconstruct any exchange’s internal premium series.
| Observation | Spot index | Perpetual midpoint | Midpoint basis | Estimated next rate, per eight hours |
|---|---|---|---|---|
| Morning | 50,000 | 50,080 | +16 bp | +0.015% |
| Afternoon | 50,500 | 50,474.75 | −5 bp | +0.008% |
The afternoon price gap is negative while the estimate remains positive. That is not sufficient evidence of a data error. Inspect the calculation window, price inputs and interest treatment first. Also check timestamps: market-data freshness can explain an apparent contradiction before a sophisticated market story is needed.
A long/short account ratio supplies another separate observation: the relative counts in a defined holder population. A majority of long accounts does not determine the price gap or the next funding payment, and it does not reveal the size or external hedges of those accounts.
Compare intervals before comparing venues
A funding percentage needs its period. A hypothetical +0.003% hourly rate and +0.016% eight-hour rate cannot be ranked by their displayed percentages alone. If the hourly rate stayed unchanged for eight settlements at constant notional, the simple total would be +0.024%. That conditional arithmetic exceeds +0.016%; it is not a forecast of either venue’s payments.
Hyperliquid’s funding documentation says its funding is paid hourly and distinguishes an eight-hour formula rate from the hourly payment fraction. It also uses the oracle price to value funding payments, rather than the mark. Read the actual field definition before applying any conversion.
Do not annualize a perpetual’s current price basis by borrowing the days-to-expiry formula used for dated futures: this contract has no fixed expiry. Annualizing a funding observation requires a stated persistence assumption and still does not establish a dependable yield.
A two-ledger example: spread change plus funding
A hypothetical trader buys 0.2 BTC spot at 50,000 and shorts an equivalent linear perpetual at 50,100. Later, both legs close: spot sells at 51,000 and the perpetual short buys back at 51,020. Assume both legs fill completely at those prices and settle in the same quote unit.
| Component | Calculation | Result, USDT |
|---|---|---|
| Spot price result | 0.2 × (51,000 − 50,000) | +200 |
| Perpetual price result | 0.2 × (50,100 − 51,020) | −184 |
| Combined price result | 200 − 184 | +16 |
| Invented net funding received | Sum of actual settlements assumed for this example | +2 |
| Four commissions at an invented 0.04% each | (10,000 + 10,020 + 10,200 + 10,204) × 0.0004 | −16.1696 |
| Result after these modeled items | 16 + 2 − 16.1696 | +1.8304 |
The price component comes from the gap contracting from 100 to 20 USDT per BTC: 0.2 × (100 − 20) = 16. Funding contributes a separate 2 USDT. Adding the opening 20-basis-point premium directly to a funding percentage would mix a price relationship with payments over time.
The small positive modeled result is fragile. It excludes slippage, financing, transfers and operational costs. The order-book depth worksheet helps assess the execution assumption. None of these invented values represents an available opportunity or observed return.
The price gap also does not choose an acceptable position size. If a separate plan sizes exposure from a volatility-based stop, the ATR sizing ledger shows how a wider assumed stop changes quantity and how a later amendment can expand modeled loss. That single-position calculation still needs a separate review of both legs and their collateral.
Where the apparent hedge can fail
A spot holding and an opposite perpetual position do not remove every risk. The gap can widen after entry, increasing the short’s loss relative to the spot gain. Collateral may be held elsewhere, so a profitable spot leg may not supply margin to the perpetual account in time. One leg can fill while the other remains incomplete.
Funding can also reverse while the trade remains open. There is no expiry deadline that guarantees recovery of an adverse gap. A large positive rate can accompany severe execution and liquidation risk rather than easy income.
Before interpreting either number, save the contract, price references, synchronized timestamps, rate status, settlement interval and applicable rules. Then ask two separate questions: what price relationship exists now, and what cash flow could the held position actually incur? Keeping those questions separate makes the evidence reproducible without pretending it predicts the next move.