Signals Package

Overview

The Signals package is a trading bot that optimizes and executes strategies based on market signals. It uses a natural selection algorithm to evolve trading models and supports multiple platforms.

Installation

Executable Files

Download the appropriate executable for your operating system:

• Windows: signals-windows-amd64.exe
• macOS: signals-darwin-arm64
• Linux:
  • ARM64: signals-linux-arm64
  • AMD64: signals-linux-amd64

Configuration

Before running the bot, configure the .env.local file with your preferred settings.

Example .env.local Configuration:

# OKX API Configuration
OKX_BASE_URL=https://www.okx.com
OKX_API_KEY=
OKX_API_SECRET=
OKX_API_PASSPHRASE=

# General Trading Configuration
SIGNALS_GENERATIONS=3
SIGNALS_GENERATIONS_DURATION=3600
SIGNALS_INSTRUMENT=DOGE-USDT-SWAP
SIGNALS_CANDLES=5
SIGNALS_TAKE_PROFIT=0.4
SIGNALS_STOP_LOSS=0.1
SIGNALS_LEVERAGE=50
SIGNALS_TRADE_MULTIPLIER=1
SIGNALS_COMMISSION=0.001
SIGNALS_COOLDOWN=300

# Trading Strategy Parameters
SIGNALS_RSI_UPPER_BOUND=60
SIGNALS_RSI_LOWER_BOUND=40

# Optimizer Configuration
SIGNALS_OPTIMIZER_POPULATION_SIZE=75
SIGNALS_OPTIMIZER_GENERATIONS=50
SIGNALS_OPTIMIZER_RETAIN_RATE=0.45
SIGNALS_OPTIMIZER_MUTATION_RATE=0.25
SIGNALS_OPTIMIZER_ELITE_COUNT=5

Binance Candlestick Data

By default Signals will use OKX candlestick data. To switch to Binance candlestick data use the following in your env file:

SIGNALS_NETWORK=binance
SIGNALS_INSTRUMENT=DOGEUSDT

Usage

Running the Optimizer

To run the natural selection algorithm for optimizing trading models:

./signals optimize

The optimizer will output a CSV file containing useful stats from each generation, including the best strategy (by fitness score) and its parameters from that generation. Each generation outputs to the console useful stats, including a copy and pastable set of model parameters for the best strategy in that generation.

Training a Basic Model

To train a model using the configured parameters and perform deep backtesting:

./signals train

This will output model statistics, including fitness scores and backtest results.

License

This package is provided as-is with no warranty express or implied whatsoever. Ensure you configure API keys securely and trade responsibly.