Skip to content

strategies.md

Latest commit

 

History

History
99 lines (70 loc) · 2.67 KB

strategies.md

File metadata and controls

99 lines (70 loc) · 2.67 KB

Trading Strategies

This document describes the trading strategies implemented in QuantumFlow.

Order Flow Imbalance Strategy

Overview

The primary strategy exploits temporary imbalances in the order book to predict short-term price movements. When there's significantly more volume on the bid side than the ask side, buying pressure is likely to push prices up.

Theory

Order book imbalance measures the relative difference between bid and ask volumes:

Imbalance = (Bid Volume - Ask Volume) / (Bid Volume + Ask Volume)
  • Imbalance > 0: More bid volume (bullish pressure)
  • Imbalance < 0: More ask volume (bearish pressure)

Weighted Imbalance

Simple imbalance treats all levels equally, but prices closer to the mid are more relevant. We use exponential decay:

Weighted Imbalance = Σ(volume_i * decay^i) for bids vs asks

where decay is typically 0.9, giving more weight to top-of-book levels.

Signal Generation

A signal is generated when:

  1. Threshold Exceeded: |imbalance| > 0.3
  2. Persistence: Imbalance in same direction for 3+ ticks
  3. Volatility Filter: Normalized volatility not extreme
  4. Spread Filter: Spread < 10 bps
  5. Momentum Confirmation: Price momentum aligns with imbalance

Confidence Calculation

confidence = (
    0.4 * min(abs(imbalance), 1.0) +           # Imbalance strength
    0.2 * min(abs(weighted_imbalance), 1.0) +  # Weighted strength
    0.2 * min(persistence / 10, 1.0) +         # Persistence score
    0.1 * volatility_bonus +                    # Low vol bonus
    0.1 * momentum_alignment                    # Momentum bonus
)

Position Sizing

Position size scales with:

  • Confidence level
  • Inverse of volatility (smaller in volatile markets)
  • Available balance
base_size = balance * 0.1  # 10% of balance
if low_volatility:
    size = base_size * 1.5
elif high_volatility:
    size = base_size * 0.5

Risk Management

  1. Position Limits: Max 1 BTC equivalent per symbol
  2. Drawdown Circuit Breaker: Pause at 5% drawdown
  3. Rate Limiting: Max 60 orders per minute
  4. Stop Loss: 2x ATR below entry

Performance Characteristics

  • Expected Win Rate: 50-55%
  • Average Holding Time: 1-5 minutes
  • Trades per Day: 10-50 (market dependent)
  • Target Sharpe: > 1.0

Future Strategy Ideas

Mean Reversion

  • Trade deviations from moving averages
  • Works well in ranging markets

Momentum

  • Follow strong price trends
  • Requires trend detection algorithms

Statistical Arbitrage

  • Cross-symbol correlation trading
  • Requires multiple symbol support

Machine Learning

  • Feature-based prediction models
  • Requires historical data collection