Graveyard / Funding signals / Binance perps
Funding against the trend
Does it pay to trade BTC when funding disagrees with price?
The idea
If price is up more than 2% in a week but funding is negative, shorts are paying to be wrong: go long. The mirror for shorts.
How we tested it
| Data | Binance BTCUSDT perpetual, 4-hour bars. |
|---|---|
| Period | 180 days (December 2025 to June 2026), then 365 days (June 2025 to June 2026). |
| Costs | 0.11% per round trip. |
| Holdout | The 365-day run served as the check on the 180-day result. |
| Control | No random-entry control was run for this test. |
Result
| 180 days: trades / return | 155 / +40.86% |
|---|---|
| 180 days: Sharpe | 2.36 |
| 365 days: trades / return | 101 / −17.36% |
| 365 days: Sharpe | −1.13 |
| 365 days: profit factor | 0.763 |
Why it failed
The short test sat inside a bear market. In a bull market funding is positive most of the time, so the long side rarely fires and the short side loses. Six months of good numbers were a regime, not an edge.
What would make it worth retesting
Only as a shorts-only bot that runs when the 90-day return is negative, switched on by hand.
Test id: carry-btcusdt-4h-v1. Past results of a backtest say nothing certain about the future.