A fixed-rate snapshot
Assume a long remains at 10,000 USDT notional and each interval settles at +0.01%. The long pays 1 USDT per interval: 3 USDT over three intervals, 21 USDT over seven days at three intervals per day, and 90 USDT over thirty days.
This is conditional, not predictive. Position value moves with mark price, the rate updates before settlement, and some contracts can change settlement frequency.
Sensitivity is more useful than one point
Over thirty days, average rates of +0.005%, +0.01% and +0.03% per interval produce teaching costs of 45, 90 and 270 USDT. A few basis points create a 225 USDT spread in the cumulative result.
A long-horizon page should therefore show ranges and interval counts, not a single precise current-rate-times-90 figure. Stress cases should include rate caps, frequency changes and resized positions.
The full rate-by-horizon matrix
Three average rates crossed with three holding periods, on a fixed 10,000 USDT long. Every cell is notional × rate × intervals, so the funding calculator will produce it.
Read across for the effect of time and down for the effect of rate. The bottom-right cell is worth pausing on: 270 USDT over thirty days at +0.03% is 2.7% of notional, an order of magnitude beyond any single round trip in fees. On long holds, fees are never the main cost.
| Average rate per interval | 3 intervals (~1 day) | 21 intervals (~7 days) | 90 intervals (~30 days) |
|---|---|---|---|
| +0.005% | 1.50 USDT | 10.50 USDT | 45.00 USDT |
| +0.010% | 3.00 USDT | 21.00 USDT | 90.00 USDT |
| +0.030% | 9.00 USDT | 63.00 USDT | 270.00 USDT |
Annualising is useful, but three assumptions usually fail
The conversion is annual ≈ rate per interval × intervals per day × 365. At three settlements a day, +0.01% is roughly 10.95% a year and +0.03% roughly 32.85%. That framing makes it easy to compare holding cost against other uses of the same capital.
Three assumptions behind it rarely hold: the rate does not stay fixed for a year, the position value does not stay fixed for a year, and the settlement frequency can be changed by the venue. Treating the annualised figure as an expected cost will systematically over- or under-state it depending on the rate environment you happened to convert in.
The sound use is to read it as the intensity of holding cost right now — a spot rate, not next year's realised interest. Estimating true cost still means going back to the interval-by-interval series.
Annualised holding cost ≈ rate per interval × settlements per day × 365| Rate per interval | Annualised at 3 settlements/day | What it describes |
|---|---|---|
| +0.005% | about 5.48% | Current intensity only |
| +0.010% | about 10.95% | Current intensity only |
| +0.030% | about 32.85% | Rarely sustained |
Only open positions settle
Closing before the timestamp generally avoids that interval; closing immediately after means it has already settled. Repeated exits intended to avoid funding may add taker fees and slippage.
Compare the expected value of holding through one more interval with the friction of exiting and re-entering. Funding is one component, not a tax that must always be avoided.
Operational check
Preserve the rate, next settlement time, fetch timestamp and source used in each estimate, then reconcile against the account ledger. Persistent underestimation can reveal omitted dynamic notional or frequency changes.
DZPTO uses public snapshots for scenarios. The platform settlement record remains authoritative.
Official sources and calculation boundary
Settlement frequency and payment direction follow OKX's funding FAQ. The three rates in the table are scenario inputs, not an expectation about the next thirty days — a distinction that is especially easy to lose on a long-horizon page.
Next checks in this series
Funding rates
Who pays when perpetual funding is negative: longs or shorts?
Reproducible scenario guide
Review methods
A reproducible monthly funding-rate review for BTC, ETH and SOL
Reproducible scenario guide
Liquidation and margin risk
BTC at 5x, 10x, 20x and 50x: how liquidation risk changes
Reproducible scenario guide
Frequently asked questions
Does a positive rate always mean longs pay?
In the common perpetual mechanism, positive funding is generally paid by longs to shorts, but verify the direction shown for the specific contract.
Can the current rate forecast thirty days?
No. It can only define a sensitivity scenario because rate, frequency and position value can all change.
Why not just show current rate × 90 as one number?
Because its precision would be fictional. It treats a series that moves every interval as a constant, and a single figure invites readers to treat it as a commitment. Showing a range and a sensitivity is the honest way to say this is a conditional scenario, not a forecast.
How does a changing position value affect cumulative funding?
Funding is charged on the position value at each settlement, so a long in profit pays more each interval and a long in drawdown pays less. Estimating a long horizon from a fixed notional drifts systematically in a trending market.