Intervals & normalization
Venues settle funding every 1, 4 or 8 hours. Raw rates are not comparable until they are put on one clock.
The problem
Funding rates come with a period attached. Most venues settle every 8 hours, but many contracts run 4-hour or 1-hour intervals — and the same coin can settle on different clocks on different venues. HTX alone runs dozens of 4h and 1h contracts next to its 8h ones.
Raw rates mislead: +0.05% per 1h is a far higher cost than +0.10% per 8h, even though the second number looks bigger. Comparing them directly gets the direction of the trade wrong.
The normalization
ARBHUB converts every rate to an 8-hour equivalent before any comparison:
fr8 = native rate × (8 / native interval)- +0.05% per 1h -> +0.400% /8h
- -0.105% per 4h -> -0.210% /8h
- +0.320% per 8h -> +0.320% /8h (unchanged)
Every cell in the matrix, every spread, and every ranking uses fr8. The native rate and its interval stay visible — each cell carries a 1h / 4h / 8h badge, and the hover popover shows the as-published number.
The NORM toolbar control re-expresses the grid in other windows (1H, 4H, APR) — it rescales the same normalized rate, it never mixes clocks.
Cash flow still follows the native clock
Normalization is for comparison only. Money moves at the native cadence:
- A 1h contract at +0.05% pays 0.05% every hour — eight payments per 8h window, not one payment of 0.4%.
- A pair with a 4h leg and an 8h leg settles at different times. For part of the window you have collected one leg but not the other, and either rate can change before its next settlement.
The Spread page shows a per-leg countdown and interval (L 1h · S 8h) so the schedule is explicit before entry.
Predicted vs current rate
Venues publish the current interval’s rate and, for the next interval, a running prediction that moves with the premium until it locks at settlement. ARBHUB’s matrix uses the current rate; treat any figure for the next window as an estimate until the venue fixes it.