Claim: When perp funding rate is in the extreme positive tail (e.g., top 1–5% of its 30–90d distribution), the next 6–12 hours of spot returns are negative on average (mean reversion).

Rationale: Extreme positive funding indicates overcrowded long positioning. As funding cost spikes, marginal longs de-risk, and short-term liquidity becomes one-sided, increasing vulnerability to squeeze/reversal.

Test:

1. Define extreme funding as funding rate > P95 (or P99) of trailing 90d for each market.
2. For each extreme event, compute forward spot returns at +6h and +12h.
3. Compare to baseline unconditional returns and to non-extreme periods.

Metrics:

• Mean forward return (t-test / bootstrap CI)
• Hit rate (P[return < 0])
• Effect size vs. baseline

Falsifier: If forward returns are not significantly negative (or are positive) after extreme funding spikes, the hypothesis fails.

Notes: Control for volatility regime and major news events to avoid confounding.