test_skew_ratio.py

"""Test script to verify skew ratio calculation logic"""

import sys
import os
sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))

def calculate_skew_ratio(buy_qty: float, sell_qty: float, epsilon: float = 0.001) -> tuple:
    """Calculate skew ratio with detailed output
    
    Args:
        buy_qty: Buy volume
        sell_qty: Sell volume
        epsilon: Minimum threshold
        
    Returns:
        Tuple of (ratio, dominant_side, calculation_details)
    """
    max_side = max(buy_qty, sell_qty)
    min_side = min(buy_qty, sell_qty)
    
    # Check if both are too small
    if max_side < epsilon:
        return None, None, "Both sides below threshold"
    
    # Adjust min_side if too small
    original_min = min_side
    if min_side < epsilon:
        min_side = epsilon
        adjusted = True
    else:
        adjusted = False
    
    # Calculate ratio
    ratio = max_side / min_side
    
    # Determine dominant side
    dominant_side = "BUY" if buy_qty > sell_qty else "SELL"
    
    # Build details
    details = {
        'max_side': max_side,
        'min_side_original': original_min,
        'min_side_used': min_side,
        'adjusted': adjusted,
        'ratio': ratio,
        'dominant_side': dominant_side,
        'buy_qty': buy_qty,
        'sell_qty': sell_qty
    }
    
    return ratio, dominant_side, details

def test_scenarios():
    """Test various scenarios for skew ratio calculation"""
    
    print("=" * 60)
    print("SKEW RATIO CALCULATION TESTS")
    print("=" * 60)
    
    test_cases = [
        # (buy_qty, sell_qty, description)
        (10.0, 2.0, "Normal case - buy dominant"),
        (2.0, 10.0, "Normal case - sell dominant"),
        (10.0, 0.0, "One side zero - buy only"),
        (0.0, 10.0, "One side zero - sell only"),
        (0.0005, 10.0, "One side very small"),
        (10.0, 0.0005, "Other side very small"),
        (0.0005, 0.0005, "Both sides below threshold"),
        (100.0, 0.01, "Large imbalance"),
        (1000.0, 0.001, "Extreme imbalance"),
        (1000.0, 0.0001, "Ultra extreme imbalance"),
        (5.0, 5.0, "Perfect balance"),
        (5.0, 2.5, "2:1 ratio"),
        (7.5, 2.5, "3:1 ratio"),
    ]
    
    for buy_qty, sell_qty, description in test_cases:
        print(f"\nTest: {description}")
        print(f"Buy: {buy_qty:.4f}, Sell: {sell_qty:.4f}")
        
        ratio, side, details = calculate_skew_ratio(buy_qty, sell_qty)
        
        if ratio is None:
            print(f"Result: {details}")
        else:
            print(f"Ratio: {ratio:.2f}")
            print(f"Dominant side: {side}")
            if details['adjusted']:
                print(f"Note: Min side adjusted from {details['min_side_original']:.4f} to {details['min_side_used']:.4f}")
            
            # Check if ratio seems reasonable
            if ratio > 10000:
                print(f"WARNING: Very high ratio detected!")
                print(f"   Max side: {details['max_side']:.4f}")
                print(f"   Min side used: {details['min_side_used']:.4f}")
                print(f"   Calculation: {details['max_side']:.4f} / {details['min_side_used']:.4f} = {ratio:.2f}")

def test_real_scenario():
    """Test with real-world values that might occur"""
    
    print("\n" + "=" * 60)
    print("REAL-WORLD SCENARIO TESTS")
    print("=" * 60)
    
    # Simulate a large market buy
    print("\n1. Large Market Buy Event (100 BTC buy vs 0.1 BTC sell)")
    ratio, side, details = calculate_skew_ratio(100.0, 0.1)
    print(f"   Ratio: {ratio:.2f} - {side}")
    print(f"   This is a {ratio:.0f}:1 imbalance")
    
    # Simulate extreme one-sided action
    print("\n2. Extreme One-Sided Action (500 BTC buy vs 0 sell)")
    ratio, side, details = calculate_skew_ratio(500.0, 0.0)
    print(f"   Ratio: {ratio:.2f} - {side}")
    print(f"   Note: {details['adjusted']} adjustment applied")
    
    # Normal volatile period
    print("\n3. Normal Volatile Period (50 BTC buy vs 20 BTC sell)")
    ratio, side, details = calculate_skew_ratio(50.0, 20.0)
    print(f"   Ratio: {ratio:.2f} - {side}")
    print(f"   This is a {ratio:.1f}:1 imbalance")

def analyze_epsilon_impact():
    """Analyze impact of different epsilon values"""
    
    print("\n" + "=" * 60)
    print("EPSILON VALUE IMPACT ANALYSIS")
    print("=" * 60)
    
    epsilons = [1e-12, 1e-6, 0.0001, 0.001, 0.01, 0.1]
    test_case = (100.0, 0.0)  # 100 BTC buy, 0 sell
    
    print(f"\nTest case: Buy={test_case[0]}, Sell={test_case[1]}")
    print("\nEpsilon Value | Resulting Ratio")
    print("-" * 35)
    
    for eps in epsilons:
        ratio, _, _ = calculate_skew_ratio(test_case[0], test_case[1], epsilon=eps)
        if ratio:
            print(f"{eps:12.2e} | {ratio:15,.2f}")
        else:
            print(f"{eps:12.2e} | None")

if __name__ == "__main__":
    test_scenarios()
    test_real_scenario()
    analyze_epsilon_impact()
    
    print("\n" + "=" * 60)
    print("CONCLUSION:")
    print("- Epsilon=0.001 provides reasonable protection against division issues")
    print("- High ratios (>1000) can occur naturally with one-sided markets")
    print("- Current logic appears mathematically correct")
    print("- No artificial cap needed if we accept extreme market conditions")
    print("=" * 60)