Auto-discovered strategy
Symbol: BTC | Exchange: Bitfinex | Role: carry
Click a year to view chart
| Year | Return | Win Rate | Trades | Max DD | Sharpe |
|---|---|---|---|---|---|
| 2020 | -4.5% | 0.0% | 4 | 4.5% | -2.62 |
| 2021 | -0.1% | 40.0% | 15 | 4.0% | -0.02 |
| 2022 | -7.9% | 28.6% | 14 | 9.6% | -1.89 |
| 2023 | +9.3% | 47.3% | 91 | 7.6% | 0.71 |
| 2024 | +4.4% | 43.9% | 82 | 13.0% | 0.35 |
| 2025 | +13.5% | 61.2% | 67 | 4.4% | 1.98 |
| Window | Train Period | Val Period | Val Return | Val | Test Period | Test Return | Status |
|---|---|---|---|---|---|---|---|
| WF-1 | 2024-01→2025-06 | 2025-07→2025-12 | +0.2% | OK | 2026-01→ongoing | +0.0% | PASS |
Not yet reviewed. Run: ./review_strategy.sh funding_rate_carry
"""
Strategy: funding_rate_carry
============================
Carry-style strategy harvesting extreme funding rate premiums on BTC.
Key concept: High funding rates indicate strong bullish sentiment and
speculator positioning. When funding is at extreme levels (top 10%),
it signals strong momentum that tends to persist short-term.
We enter long during these extreme funding periods with strict trend
confirmation and tight risk management, capturing price momentum while
maintaining bounded risk appropriate for a carry strategy.
Entry conditions:
- Funding rate >= 90th percentile (100-bar lookback) AND >= 40% APR
- Strong uptrend confirmed: EMA20 > EMA50 > EMA200
- All three EMAs must be valid (no entry during warmup)
Exit conditions:
- Funding drops below median (50th percentile) - sentiment normalizing
- Short-term trend break: Close < EMA20
- Max hold: 10 bars (carry = short duration exposure)
- Stop loss: 2% (tight for carry role)
- Take profit: 5%
Role: carry
Validation gates: max_loss=-5%, max_dd=15%, min_sharpe=-0.2, min_trades=3
Expected behavior: Bounded losses in bear markets, consistent small gains
Train results (2024-01-01 to 2025-06-30):
- 2024: +4.4%, 82 trades, 44% WR, 13.0% DD
- 2025H1: +8.2%, 14 trades, 71% WR, 1.7% DD
- Total: +12.5%, Max DD 13.0%
"""
import sys
sys.path.insert(0, '/root/trade_rules')
from lib import ema
def init_strategy():
return {
'name': 'funding_rate_carry',
'role': 'carry',
'warmup': 200,
'subscriptions': [
{'symbol': 'tBTCUSD', 'exchange': 'bitfinex', 'timeframe': '4h'},
],
'parameters': {
'funding_pctl_entry': 0.90, # Top 10% funding to enter
'funding_min': 0.40, # At least 40% APR
'funding_pctl_exit': 0.50, # Exit when below median
'lookback': 100, # Percentile lookback period
'max_hold': 10, # Max bars to hold
'stop_loss': 2.0, # 2% stop loss (tight for carry)
'take_profit': 5.0, # 5% take profit
}
}
def process_time_step(ctx):
key = ('tBTCUSD', 'bitfinex')
bars = ctx['bars'].get(key, [])
i = ctx['i']
positions = ctx['positions']
params = ctx['parameters']
state = ctx['state']
# Warmup period - now handled by framework via 'warmup' field
# if i < 200 or i >= len(bars):
# return []
if i >= len(bars):
return []
funding = bars[i].funding_apr
close = bars[i].close
# Track funding history for percentile calculation
if 'funding_hist' not in state:
state['funding_hist'] = []
state['funding_hist'].append(funding)
if len(state['funding_hist']) > params['lookback']:
state['funding_hist'] = state['funding_hist'][-params['lookback']:]
# Calculate funding percentile
hist_len = len(state['funding_hist'])
if hist_len >= params['lookback']:
below = sum(1 for f in state['funding_hist'][:-1] if f < funding)
funding_pctl = below / (hist_len - 1)
else:
funding_pctl = 0.5 # Default to median if not enough history
# Calculate EMAs for trend confirmation
closes = [b.close for b in bars[:i+1]]
ema20_vals = ema(closes, 20)
ema50_vals = ema(closes, 50)
ema200_vals = ema(closes, 200)
ema20 = ema20_vals[i] if ema20_vals[i] else 0
ema50 = ema50_vals[i] if ema50_vals[i] else 0
ema200 = ema200_vals[i] if ema200_vals[i] else 0
actions = []
has_position = key in positions
if not has_position:
# Entry conditions
# 1. Strong uptrend: EMA20 > EMA50 > EMA200 (all positive)
strong_uptrend = ema20 > ema50 > ema200 > 0
# 2. Extreme funding: in top percentile AND above minimum APR
extreme_funding = (
funding_pctl >= params['funding_pctl_entry'] and
funding >= params['funding_min']
)
if strong_uptrend and extreme_funding:
actions.append({
'action': 'open_long',
'symbol': 'tBTCUSD',
'exchange': 'bitfinex',
'size': 1.0,
'stop_loss_pct': params['stop_loss'],
'take_profit_pct': params['take_profit'],
})
else:
# Exit conditions
pos = positions[key]
bars_held = i - pos.entry_bar
should_exit = False
# Exit 1: Funding normalizes (drops below median)
if funding_pctl < params['funding_pctl_exit']:
should_exit = True
# Exit 2: Short-term trend break (close below EMA20)
elif ema20 > 0 and close < ema20:
should_exit = True
# Exit 3: Max hold period reached
elif bars_held >= params['max_hold']:
should_exit = True
if should_exit:
actions.append({
'action': 'close_long',
'symbol': 'tBTCUSD',
'exchange': 'bitfinex',
})
return actions
if __name__ == '__main__':
from strategy import backtest_strategy
print("Backtesting funding_rate_carry strategy...")
results, profitable, _ = backtest_strategy(init_strategy, process_time_step)
print(f"\nSummary:")
print(f" Profitable years: {profitable}/2")
total_ret = sum(r['return'] for r in results.values())
max_dd = max(r['max_dd'] for r in results.values())
avg_sharpe = sum(r['sharpe'] for r in results.values()) / len(results)
print(f" Total return: {total_ret:.1f}%")
print(f" Max drawdown: {max_dd:.1f}%")
print(f" Avg Sharpe: {avg_sharpe:.2f}")