Skip to content

Trading Strategies

dgbit ships four registered strategies and a framework for adding more.

The base strategy class (BaseStrategy) and WaveletReversalStrategy live in dgbit_core.trading.strategy. The three example strategies (MACrossoverStrategy, RSIStrategy, BollingerBandStrategy) live in dgbit_core.trading.examples and are re-exported from dgbit_core.trading.

Built-in Strategies

Wavelet Reversal Strategy

Registry name: wavelet_reversal. generate_signal() delegates to dgbit_core.models.predictor.PricePredictor, which performs a Daubechies wavelet decomposition.

from dgbit_core.trading.strategy import WaveletReversalStrategy

# Constructor signature: (min_signal_threshold, take_profit_pct, stop_loss_pct, **kwargs)
strategy = WaveletReversalStrategy(
    min_signal_threshold=0.75,
    take_profit_pct=0.002,
    stop_loss_pct=0.005,
)

Extra keyword arguments are stored on self._extra_params but are not consumed by the strategy itself.

MA Crossover Strategy

Registry name: ma_crossover. Uses SMA, EMA, or WMA depending on ma_type (default "sma"). The signal is a normalised, clamped function of the percentage gap between fast and slow MAs (above 0.5 = fast above slow).

from dgbit_core.trading import MACrossoverStrategy

strategy = MACrossoverStrategy(
    fast_period=10,
    slow_period=20,
    ma_type="ema",   # one of: "sma", "ema", "wma"
)

RSI Strategy

Registry name: rsi. Returns 0.0 when RSI <= oversold, 1.0 when RSI >= overbought, and a linear interpolation between them otherwise.

from dgbit_core.trading import RSIStrategy

strategy = RSIStrategy(
    period=14,
    oversold=30.0,
    overbought=70.0,
)

Bollinger Bands Strategy

Registry name: bollinger_bands. Computes the close price's relative position within a period-SMA Bollinger Band (with std_dev standard deviations). The signal is the clamped ratio (close - lower) / (upper - lower). There is no separate breakout mode in the current implementation.

from dgbit_core.trading import BollingerBandStrategy

strategy = BollingerBandStrategy(
    period=20,
    std_dev=2.0,
)

Using Strategies

With Backtesting

from dgbit_core.backtesting import Backtester, BacktestConfig
from dgbit_core.trading.strategy import WaveletReversalStrategy
from dgbit_core.data.data_fetcher import BybitDataFetcher

fetcher = BybitDataFetcher(api_key="", api_secret="", testnet=True)
data = fetcher.get_kline_data("BTCUSDT", interval="15", limit=1000)

config = BacktestConfig(initial_capital=10000.0)
backtester = Backtester(config=config)
backtester.strategy = WaveletReversalStrategy()

result = backtester.run(data)
print(f"Win Rate: {result.metrics['win_rate']:.2%}")

With the API

The /api/strategies/{strategy_name}/signal endpoint accepts only a symbol query parameter and proxies the request to the strategy service over NNG:

import httpx

response = httpx.post(
    "http://localhost:8000/api/strategies/wavelet_reversal/signal",
    params={"symbol": "BTCUSDT"},
)
print(response.json())

The response shape is whatever the strategy service returns over the bus; consult dgbit_services.strategy for the current schema.

Strategy Registry

All strategies are registered in the global registry:

from dgbit_core.trading.strategy import strategy_registry

# List all available strategies
strategies = strategy_registry.list_strategies()
for name, metadata in strategies.items():
    print(f"{name}: {metadata.description}")

# Create a strategy by name
strategy = strategy_registry.create(
    "wavelet_reversal",
    min_signal_threshold=0.8
)

Strategy Comparison

Strategy SignalType Default direction Notes
Wavelet Reversal REVERSAL LONG Wraps PricePredictor
MA Crossover MOMENTUM LONG Configurable MA type
RSI MEAN_REVERSION LONG Returns 0/1 at thresholds
Bollinger Bands BREAKOUT BOTH Reports position within bands

Combining Strategies

You can combine signals from multiple strategies manually:

from dgbit_core.trading import WaveletReversalStrategy, RSIStrategy

wavelet = WaveletReversalStrategy()
rsi = RSIStrategy()

wavelet_signal = wavelet.generate_signal(data)
rsi_signal = rsi.generate_signal(data)

# Note: RSI emits 0.0 at oversold and 1.0 at overbought, so the
# directional meaning of "agreement" depends on the strategy semantics.
combined_signal = (wavelet_signal + rsi_signal) / 2

Next Steps