Methodology
The funding rate is the periodic payment between longs and shorts on a perpetual swap, paid every 8 hours. It is a direct, exchange-settled measure of positioning pressure: persistently positive funding means longs are paying to stay long (crowded longs), persistently negative means shorts are paying (crowded shorts). The z-score standardizes the raw rate against its own recent history, so "4.5" means the current funding rate is 4.5 standard deviations above the norm.

Calculation

Each UTC day's funding payments (3 per day on BitMEX XBTUSD) are averaged into a single daily funding rate (in % per 8-hour interval). The z-score at day t is (f_t − μ)/σ, where μ and σ are the mean and sample standard deviation of the strictly prior 90 days (min 30 prior observations required before any z is produced). The current day is never included in its own benchmark; warm-up periods are left blank, never fabricated. A 90-day window is the standard benchmark for funding z-scores (short enough to catch positioning shifts, long enough to be stable); |z| is capped at 10 for display so that the handful of extreme crash days (which reach raw values beyond −20) do not crush the scale — the axis runs ±10 like the reference chart.

Reading the chart

The dashed red line marks +3σ: funding more than 3 standard deviations above its 90-day norm — a strong signal of excess long positioning (↑ Longs Excess). The dashed green line marks −3σ: excess short positioning (↓ Shorts Excess). These are statistical extremes, not hard trade triggers; they identify when positioning is unusually stretched in one direction.

Data and caveats

Funding history comes from the BitMEX XBTUSD perpetual swap, the longest-running BTC perp (listed May 2016) and the dominant funding venue during 2016–2019. Binance and Bybit APIs are geo-blocked from this server and OKX's public history endpoint only serves ~2 months, so a single-venue series is the only complete keyless source — and the most consistent one (no venue-splice regime shifts). Funding is a venue-specific rate: other exchanges quote similar but not identical rates, and the z-score's recent reading will differ from aggregated multi-exchange feeds. This is a positioning gauge, not a price predictor.