Graveyard / Funding signals / Binance perps
Fading extreme funding
Does trading against extreme funding rates work on Bitcoin?
The idea
Very high funding means crowded longs. Very low funding means crowded shorts. Bet that the crowd is wrong and price snaps back.
How we tested it
| Data | Binance BTCUSDT perpetual, 1-hour bars. Signal: funding z-score over 30 days beyond ±1.5. Take profit and stop at 1.5%, 16-hour time stop. |
|---|---|
| Period | 20 December 2025 to 21 June 2026 (180 days). |
| Costs | 0.11% per round trip. |
| Holdout | No separate holdout was run. |
| Control | No random-entry control was run for this test. |
Result
| Trades | 118 |
|---|---|
| Win rate | 46.6% (break-even 53.7%) |
| Return | −13.11% |
| Profit factor | 0.819 |
| Sharpe | −1.32 |
| Max drawdown | −18.97% |
Why it failed
The market trended during the test. Crowded positioning stayed crowded and price kept going. Extremes did not revert.
What would make it worth retesting
Only with a filter for ranging markets, tested inside such a period. Not before someone shows this test period was unusual.
Test id: mr-btcusdt-1h-v1. Past results of a backtest say nothing certain about the future.