This page states every calculation rule the site follows. It is published because a number without a method is not data.
Every figure on the site carries a stated snapshot time in UTC and a named source. Where sources publish at different times, the site prints each reading with its own timestamp rather than blending them. Week-on-week comparisons are made between readings taken at the same point in the weekly cycle wherever the source permits it.
| Term | Formula | Why it is defined this way |
|---|---|---|
| Annualised funding, 8-hour settlement | rate x 3 x 365 x 100 | Three settlement periods per day, 365 days a year. A 0.005% per-8h print reads as about 5.47% annualised, not 0.5%. |
| Annualised funding, 1-hour settlement | rate x 24 x 365 x 100 | Twenty-four settlement periods per day. Hyperliquid settles hourly; this clock matters whenever an hourly reading is quoted. |
| Open interest | sum of the notional value of open contracts, USD | USD-denominated throughout. Where two sources report different measures for the same venue - a one-sided tracker cohort and the venue's own two-sided count - both are printed with their cohorts named. |
| Liquidation windows | rolling 24-hour totals as published; windows are never summed into a weekly total | Overlapping windows double-count. A weekly figure computed by adding daily windows would overstate the week. |
| Volume-to-open-interest ratio | 30-day volume / open interest | Separates venues that hold committed capital from venues that cycle it. High ratios usually indicate incentive-driven churn. |
| Week-on-week change | absolute change and percentage change, always both | A percentage alone hides small-base distortion; an absolute change alone hides scale. |
| Unavailable data | "not disclosed" | We never interpolate, backfill or carry a stale value forward. If a figure cannot be verified against a whitelisted source, it does not ship. |
Market share figures are meaningless without the denominator. The site names the cohort next to every share number - for perp DEX open interest, the tracked cohort and basis printed side by side in the Data Report; for venue-level readings, the source's own stated cohort. Where published third-party share figures differ from ours, both are shown with the cohort named - a difference in definition is not a data error, and silently picking one is how bad comparisons get published.
Only whitelisted sources are used as the basis for a factual claim. The whitelist is published on the Sources page. A claim that cannot be traced to a whitelisted source does not ship, regardless of how widely it has been repeated elsewhere. Where a figure is attributed to a market participant rather than a regulator or exchange, the site labels it as an estimate.
Corrections are appended to the affected page with the date and the substance of the change. We do not silently edit figures. If an error changed a conclusion, the conclusion is restated and the original reading is left visible.