#!/usr/bin/env python3 """ OPTIMISEUR PRICE EXTREMES - La stratégie la plus prometteuse! PF: 19.99 (gain moyen 20x la perte!) ROI: +3.45% WR: 54% DD: -0.06% Principe: - Acheter quand prix <= percentile bas (ex: 10%) - Vendre quand prix >= percentile haut (ex: 90%) - Mean reversion basique mais TRÈS efficace On optimise: 1. Percentiles (5/95, 10/90, 15/85, etc.) 2. Lookback period (50, 100, 200 bougies) 3. Paramètres martingale """ from martingale_backtester import MartingaleBacktester from true_martingale import TrueMartingaleStrategy import numpy as np import json from typing import Optional, Tuple import pandas as pd # Créer une version customizable de PriceExtremes class OptimizedPriceExtremes(TrueMartingaleStrategy): """Version optimisée de Price Extremes avec paramètres ajustables""" def __init__(self, percentile_low=10, percentile_high=90, lookback=100, **kwargs): super().__init__(**kwargs) self.percentile_low = percentile_low self.percentile_high = percentile_high self.lookback = lookback def generate_base_signal(self, df: pd.DataFrame) -> Optional[Tuple[str, float]]: """Generate signal based on price percentiles""" current = df.iloc[-1] if len(df) < self.lookback: return None # Prix actuel vs percentiles recent_prices = df['close'].tail(self.lookback) p_low = recent_prices.quantile(self.percentile_low / 100.0) p_high = recent_prices.quantile(self.percentile_high / 100.0) # Acheter aux extrêmes bas, vendre aux extrêmes hauts if current['close'] <= p_low: return ('LONG', 1.0) # Prix très bas → acheter elif current['close'] >= p_high: return ('SHORT', 1.0) # Prix très haut → vendre return None # Prix neutre print("=" * 80) print("🎯 OPTIMISEUR PRICE EXTREMES - Stratégie PF 19.99") print("=" * 80) # Fetch data backtester = MartingaleBacktester( initial_capital=1000.0, trading_fee_pct=0.0005, timeout_minutes=60 ) print("\n📥 Téléchargement données...") df = backtester.fetch_binance_data('BTCUSDT', '1m', 1000) if df is None: exit(1) print(f"✅ {len(df)} bougies\n") # ============ GRILLE D'OPTIMISATION ============ print("🔬 PHASE 1: OPTIMISATION PERCENTILES & LOOKBACK") print("=" * 80) # Test percentiles percentile_configs = [ (5, 95), # Très extrêmes (10, 90), # Baseline (celui qui marche) (15, 85), # Moins extrêmes (20, 80), # Modérés (25, 75), # Larges ] lookback_periods = [50, 100, 150, 200] # Paramètres martingale de base (on optimisera après) base_martingale_config = { 'base_position_size': 11.0, 'max_martingale_level': 6, 'base_leverage': 20, 'max_leverage': 1000, 'stop_loss_pct': 0.003, 'take_profit_ratio': 3.0, 'min_signal_score': 0.0 } phase1_results = [] test_num = 0 total_tests = len(percentile_configs) * len(lookback_periods) for p_low, p_high in percentile_configs: for lookback in lookback_periods: test_num += 1 print(f"\r🧪 Test {test_num}/{total_tests} | P{p_low}/{p_high} LB{lookback}", end="", flush=True) try: strategy = OptimizedPriceExtremes( percentile_low=p_low, percentile_high=p_high, lookback=lookback, **base_martingale_config ) result = backtester.run_backtest(df, 'BTCUSDT', strategy, verbose=False) if result and result.total_trades > 0: phase1_results.append({ 'percentile_low': p_low, 'percentile_high': p_high, 'lookback': lookback, 'roi': result.total_return_pct, 'win_rate': result.win_rate, 'profit_factor': result.profit_factor, 'trades': result.total_trades, 'max_dd': result.max_drawdown_pct, 'capital': result.final_capital, 'avg_win': result.avg_win_pct, 'avg_loss': result.avg_loss_pct }) except: pass print("\n\n" + "=" * 80) print("🏆 TOP 10 COMBINAISONS PERCENTILES/LOOKBACK") print("=" * 80) if not phase1_results: print("\n❌ Aucun résultat en Phase 1! Arrêt.") exit(1) phase1_results.sort(key=lambda x: x['profit_factor'], reverse=True) top_10_phase1 = phase1_results[:10] print(f"\n{'#':<3} {'P-Low':>6} {'P-High':>7} {'LB':>5} {'ROI %':>8} {'WR%':>5} " f"{'PF':>7} {'Tr':>4} {'DD%':>7} {'Cap $':>10}") print("-" * 80) for rank, res in enumerate(top_10_phase1, 1): emoji = "🥇" if rank == 1 else "🥈" if rank == 2 else "🥉" if rank == 3 else f"{rank:2}" status = "✅" if res['roi'] > 0 else "❌" print(f"{emoji:<3} {res['percentile_low']:>6} {res['percentile_high']:>7} " f"{res['lookback']:>5} {res['roi']:>7.2f}% {res['win_rate']*100:>4.1f}% " f"{res['profit_factor']:>7.2f} {res['trades']:>4} " f"{res['max_dd']:>6.2f}% {res['capital']:>9.2f} {status}") # Meilleure config phase 1 best_p1 = top_10_phase1[0] print(f"\n✨ MEILLEURE COMBINAISON:") print(f" Percentiles: {best_p1['percentile_low']}% / {best_p1['percentile_high']}%") print(f" Lookback: {best_p1['lookback']} bougies") print(f" ROI: {best_p1['roi']:+.2f}%") print(f" PF: {best_p1['profit_factor']:.2f}") print(f" Trades: {best_p1['trades']}") # ============ PHASE 2: OPTIMISATION MARTINGALE ============ print("\n\n🔬 PHASE 2: OPTIMISATION PARAMÈTRES MARTINGALE") print("=" * 80) # Utiliser les meilleurs percentiles trouvés best_percentiles = { 'percentile_low': best_p1['percentile_low'], 'percentile_high': best_p1['percentile_high'], 'lookback': best_p1['lookback'] } # Grille martingale position_sizes = [3, 5, 7, 9, 11, 13, 15] max_levels = [5, 6, 7, 8] leverages = [ (10, 500), (15, 750), (20, 1000), (25, 1000), ] sl_pcts = [0.002, 0.0025, 0.003, 0.0035, 0.004] tp_ratios = [2.0, 2.5, 3.0, 3.5, 4.0] phase2_results = [] test_num = 0 # Smart sampling au lieu de tout tester num_samples = 100 print(f"🎯 Test de {num_samples} combinaisons intelligentes...\n") for i in range(num_samples): test_num += 1 # Sample random mais intelligent pos_size = np.random.choice(position_sizes) max_lv = np.random.choice(max_levels) base_lev, max_lev = leverages[np.random.randint(0, len(leverages))] sl = np.random.choice(sl_pcts) tp = np.random.choice(tp_ratios) print(f"\r🧪 Test {test_num}/{num_samples} | Meilleur ROI: " f"{max([r['roi'] for r in phase2_results], default=0):+.2f}%", end="", flush=True) try: strategy = OptimizedPriceExtremes( **best_percentiles, base_position_size=pos_size, max_martingale_level=max_lv, base_leverage=base_lev, max_leverage=max_lev, stop_loss_pct=sl, take_profit_ratio=tp, min_signal_score=0.0 ) result = backtester.run_backtest(df, 'BTCUSDT', strategy, verbose=False) if result and result.total_trades > 0: phase2_results.append({ 'config': { **best_percentiles, 'base_position_size': pos_size, 'max_martingale_level': max_lv, 'base_leverage': base_lev, 'max_leverage': max_lev, 'stop_loss_pct': sl, 'take_profit_ratio': tp }, 'roi': result.total_return_pct, 'win_rate': result.win_rate, 'profit_factor': result.profit_factor, 'trades': result.total_trades, 'max_dd': result.max_drawdown_pct, 'capital': result.final_capital, 'avg_win': result.avg_win_pct, 'avg_loss': result.avg_loss_pct, 'max_consec_losses': result.max_consecutive_losses }) except: pass print("\n\n" + "=" * 80) print("🏆 TOP 20 CONFIGURATIONS COMPLÈTES") print("=" * 80) phase2_results.sort(key=lambda x: x['roi'], reverse=True) top_20 = phase2_results[:20] print(f"\n{'#':<3} {'Pos$':>5} {'Lv':>3} {'Leg':>8} {'SL%':>5} {'TP':>4} " f"{'ROI%':>7} {'WR%':>5} {'PF':>7} {'DD%':>7} {'Tr':>3}") print("-" * 85) for rank, res in enumerate(top_20, 1): cfg = res['config'] emoji = "🥇" if rank == 1 else "🥈" if rank == 2 else "🥉" if rank == 3 else f"{rank:2}" status = "✅" if res['roi'] > 0 else "❌" lev_str = f"{cfg['base_leverage']}→{cfg['max_leverage']}" print(f"{emoji:<3} {cfg['base_position_size']:>5.0f} {cfg['max_martingale_level']:>3} " f"{lev_str:>8} {cfg['stop_loss_pct']*100:>4.2f} {cfg['take_profit_ratio']:>4.1f} " f"{res['roi']:>6.2f}% {res['win_rate']*100:>4.1f}% " f"{res['profit_factor']:>7.2f} {res['max_dd']:>6.2f}% {res['trades']:>3} {status}") # Meilleur résultat final best = top_20[0] print(f"\n\n{'=' * 80}") print(f"🏆 CONFIGURATION OPTIMALE PRICE EXTREMES") print(f"{'=' * 80}") print(f"\n📊 PERFORMANCE:") print(f" ROI: {best['roi']:+.2f}% {'✅ PROFITABLE!' if best['roi'] > 0 else '❌'}") print(f" Capital final: ${best['capital']:.2f}") print(f" Gain: ${best['capital'] - 1000:.2f}") print(f" Win Rate: {best['win_rate']*100:.1f}%") print(f" Profit Factor: {best['profit_factor']:.2f} 🚀") print(f" Max Drawdown: {best['max_dd']:.2f}%") print(f" Trades: {best['trades']}") print(f" Max pertes consec: {best['max_consec_losses']}") if best['trades'] > 0: print(f"\n💰 STATISTIQUES TRADES:") print(f" Gain moyen: +{best['avg_win']:.2f}%") print(f" Perte moyenne: {best['avg_loss']:.2f}%") ratio = abs(best['avg_win'] / best['avg_loss']) if best['avg_loss'] != 0 else 0 print(f" Ratio Gain/Perte: {ratio:.2f}x") cfg = best['config'] print(f"\n⚙️ PARAMÈTRES OPTIMAUX:") print(f"\n 📐 PERCENTILES:") print(f" Percentile BAS: {cfg['percentile_low']}% (LONG si prix <= P{cfg['percentile_low']})") print(f" Percentile HAUT: {cfg['percentile_high']}% (SHORT si prix >= P{cfg['percentile_high']})") print(f" Lookback period: {cfg['lookback']} bougies") print(f"\n 💰 MARTINGALE:") print(f" Position de base: ${cfg['base_position_size']:.2f}") print(f" Max level: {cfg['max_martingale_level']} (x{2**cfg['max_martingale_level']})") print(f" Leverage: {cfg['base_leverage']}x → {cfg['max_leverage']}x") print(f" Stop Loss: {cfg['stop_loss_pct']*100:.3f}%") print(f" Take Profit: {cfg['take_profit_ratio']:.1f}x SL") # Analyse mathématique max_position = cfg['base_position_size'] * (2 ** cfg['max_martingale_level']) print(f"\n🧮 ANALYSE MATHÉMATIQUE:") print(f" Position max: ${max_position:.2f}") print(f" Marge sécurité: {(1000 / max_position):.2f}x") print(f" Proba ruine: {(0.5 ** (cfg['max_martingale_level'] + 1)) * 100:.4f}%") print(f"{'=' * 80}") # Code Python print(f"\n📝 CODE POUR UTILISER:") print("=" * 80) print(f""" from optimize_price_extremes import OptimizedPriceExtremes strategy = OptimizedPriceExtremes( # Percentiles percentile_low={cfg['percentile_low']}, percentile_high={cfg['percentile_high']}, lookback={cfg['lookback']}, # Martingale base_position_size={cfg['base_position_size']}, max_martingale_level={cfg['max_martingale_level']}, base_leverage={cfg['base_leverage']}, max_leverage={cfg['max_leverage']}, stop_loss_pct={cfg['stop_loss_pct']}, take_profit_ratio={cfg['take_profit_ratio']}, min_signal_score=0.0 ) # STATISTIQUES: # ROI: {best['roi']:+.2f}% # Win Rate: {best['win_rate']*100:.1f}% # Profit Factor: {best['profit_factor']:.2f} # Max DD: {best['max_dd']:.2f}% # LOGIQUE: # LONG si prix <= percentile {cfg['percentile_low']}% des {cfg['lookback']} dernières bougies # SHORT si prix >= percentile {cfg['percentile_high']}% des {cfg['lookback']} dernières bougies # Sinon: rien (on attend les extrêmes) """) print("=" * 80) # Sauvegarder with open('price_extremes_optimized.json', 'w') as f: json.dump({ 'best_config': best['config'], 'best_metrics': best, 'top_20': top_20, 'phase1_results': phase1_results, 'phase2_results': phase2_results }, f, indent=2, default=str) print(f"\n💾 Résultats: price_extremes_optimized.json") # Stats globales profitable = [r for r in phase2_results if r['roi'] > 0] print(f"\n📈 STATISTIQUES GLOBALES:") print(f" Configs testées: {len(phase2_results)}") print(f" Configs profitables: {len(profitable)} ({len(profitable)/len(phase2_results)*100:.1f}%)") if profitable: avg_roi = np.mean([r['roi'] for r in profitable]) avg_pf = np.mean([r['profit_factor'] for r in profitable]) print(f" ROI moyen: +{avg_roi:.2f}%") print(f" PF moyen: {avg_pf:.2f}") print(f"\n{'=' * 80}") print("🎉 OPTIMISATION TERMINÉE!") print(f"{'=' * 80}")