#!/usr/bin/env python3 """ BetPredictAI - Historical Betting Slips Generator Genera bollette storiche basate sui 400 match di verifica Calcola profitto/perdita reale basato sulle quote simulate """ import json import os from datetime import datetime, timedelta from collections import defaultdict import random BASE_DIR = os.path.dirname(os.path.abspath(__file__)) DATA_DIR = os.path.join(BASE_DIR, '..', 'data') # League flags LEAGUE_FLAGS = { 'E0': '🏴󠁧󠁢󠁥󠁮󠁧󠁿', 'D1': '🇩🇪', 'I1': '🇮🇹', 'SP1': '🇪🇸', 'F1': '🇫🇷' } LEAGUE_NAMES = { 'E0': 'Premier League', 'D1': 'Bundesliga', 'I1': 'Serie A', 'SP1': 'La Liga', 'F1': 'Ligue 1' } def calculate_odds(prediction, home_prob, draw_prob, away_prob, correct): """ Calcola quote realistiche basate sulle probabilità Quote più alte per eventi meno probabili """ prob_map = { 'HOME': home_prob / 100, 'DRAW': draw_prob / 100, 'AWAY': away_prob / 100 } prob = prob_map.get(prediction, 0.33) # Calcola quota base (1/prob) con margine bookmaker 5-10% if prob > 0: base_odds = 1 / prob margin = random.uniform(0.92, 0.96) # Margine bookmaker odds = round(base_odds * margin, 2) odds = max(1.10, min(odds, 15.00)) # Limiti realistici else: odds = 3.00 return odds def group_matches_by_date(results): """Raggruppa i match per data""" by_date = defaultdict(list) for match in results: by_date[match['date']].append(match) return dict(sorted(by_date.items())) def calculate_slip_stake(): """Puntata standard per ogni bolletta""" return 10.0 # €10 per bolletta def generate_daily_slips(matches_by_date, min_confidence=55, max_selections=4): """ Genera bollette giornaliere - Seleziona solo match con confidence >= min_confidence - Massimo max_selections per bolletta - Priorità ai match con confidence più alta """ slips = [] slip_id = 1 for date, matches in matches_by_date.items(): # Filtra match con confidence sufficiente high_conf_matches = [m for m in matches if m['confidence'] >= min_confidence] if len(high_conf_matches) < 2: continue # Skip giorni con meno di 2 match validi # Ordina per confidence decrescente high_conf_matches.sort(key=lambda x: x['confidence'], reverse=True) # Prendi i migliori match (max_selections) selected = high_conf_matches[:max_selections] # Crea bolletta slip_matches = [] total_odds = 1.0 all_correct = True for match in selected: odds = calculate_odds( match['predicted'], match['home_prob'], match['draw_prob'], match['away_prob'], match['correct'] ) total_odds *= odds if not match['correct']: all_correct = False slip_matches.append({ 'id': f"h_{match['home_team']}_{match['away_team']}_{match['date']}", 'home_team': match['home_team'], 'away_team': match['away_team'], 'league': match['league'], 'league_name': match['league_name'], 'flag': LEAGUE_FLAGS.get(match['league'], '⚽'), 'date': match['date'], 'time': match.get('time', '15:00'), 'prediction': match['predicted'], 'predicted_sign': '1' if match['predicted'] == 'HOME' else ('X' if match['predicted'] == 'DRAW' else '2'), 'actual_result': match['actual'], 'actual_sign': '1' if match['actual'] == 'HOME' else ('X' if match['actual'] == 'DRAW' else '2'), 'home_goals': match['home_goals'], 'away_goals': match['away_goals'], 'confidence': match['confidence'], 'odds': odds, 'is_correct': match['correct'], 'home_prob': match['home_prob'], 'draw_prob': match['draw_prob'], 'away_prob': match['away_prob'] }) stake = calculate_slip_stake() potential_win = round(stake * total_odds, 2) actual_win = potential_win if all_correct else 0 profit = actual_win - stake # Calcola probabilità combinata win_probability = 1.0 for m in slip_matches: prob_map = {'HOME': m['home_prob'], 'DRAW': m['draw_prob'], 'AWAY': m['away_prob']} win_probability *= prob_map[m['prediction']] / 100 slip = { 'id': f"SLIP-{date.replace('-', '')}-{slip_id:03d}", 'date': date, 'created_at': f"{date}T12:00:00", 'strategy': f"AI High Confidence (>={min_confidence}%)", 'matches': slip_matches, 'total_odds': round(total_odds, 2), 'stake': stake, 'potential_win': potential_win, 'win_probability': round(win_probability * 100, 2), 'status': 'won' if all_correct else 'lost', 'actual_win': actual_win, 'profit': profit, 'correct_predictions': sum(1 for m in slip_matches if m['is_correct']), 'total_predictions': len(slip_matches) } slips.append(slip) slip_id += 1 return slips def calculate_statistics(slips): """Calcola statistiche aggregate""" total_slips = len(slips) won_slips = sum(1 for s in slips if s['status'] == 'won') lost_slips = sum(1 for s in slips if s['status'] == 'lost') total_staked = sum(s['stake'] for s in slips) total_won = sum(s['actual_win'] for s in slips) total_profit = sum(s['profit'] for s in slips) win_rate = (won_slips / total_slips * 100) if total_slips > 0 else 0 roi = (total_profit / total_staked * 100) if total_staked > 0 else 0 # Calcola streak current_streak = 0 max_win_streak = 0 max_loss_streak = 0 temp_streak = 0 last_status = None for slip in sorted(slips, key=lambda x: x['date']): if slip['status'] == last_status: temp_streak += 1 else: if last_status == 'won': max_win_streak = max(max_win_streak, temp_streak) elif last_status == 'lost': max_loss_streak = max(max_loss_streak, temp_streak) temp_streak = 1 last_status = slip['status'] # Final streak if last_status == 'won': max_win_streak = max(max_win_streak, temp_streak) current_streak = temp_streak elif last_status == 'lost': max_loss_streak = max(max_loss_streak, temp_streak) current_streak = -temp_streak # Statistiche per lega by_league = defaultdict(lambda: {'won': 0, 'lost': 0, 'profit': 0, 'staked': 0}) for slip in slips: for match in slip['matches']: league = match['league'] by_league[league]['staked'] += slip['stake'] / len(slip['matches']) if match['is_correct']: by_league[league]['won'] += 1 else: by_league[league]['lost'] += 1 league_stats = {} for league, stats in by_league.items(): total = stats['won'] + stats['lost'] league_stats[league] = { 'name': LEAGUE_NAMES.get(league, league), 'flag': LEAGUE_FLAGS.get(league, '⚽'), 'matches': total, 'won': stats['won'], 'lost': stats['lost'], 'accuracy': round(stats['won'] / total * 100, 1) if total > 0 else 0 } # Best e worst days sorted_by_profit = sorted(slips, key=lambda x: x['profit'], reverse=True) best_days = sorted_by_profit[:5] worst_days = sorted_by_profit[-5:] # Monthly breakdown monthly = defaultdict(lambda: {'slips': 0, 'won': 0, 'staked': 0, 'profit': 0}) for slip in slips: month = slip['date'][:7] # YYYY-MM monthly[month]['slips'] += 1 monthly[month]['staked'] += slip['stake'] monthly[month]['profit'] += slip['profit'] if slip['status'] == 'won': monthly[month]['won'] += 1 monthly_stats = [] for month, stats in sorted(monthly.items()): monthly_stats.append({ 'month': month, 'slips': stats['slips'], 'won': stats['won'], 'win_rate': round(stats['won'] / stats['slips'] * 100, 1) if stats['slips'] > 0 else 0, 'staked': round(stats['staked'], 2), 'profit': round(stats['profit'], 2), 'roi': round(stats['profit'] / stats['staked'] * 100, 1) if stats['staked'] > 0 else 0 }) return { 'total_slips': total_slips, 'won_slips': won_slips, 'lost_slips': lost_slips, 'win_rate': round(win_rate, 1), 'total_staked': round(total_staked, 2), 'total_won': round(total_won, 2), 'total_profit': round(total_profit, 2), 'roi': round(roi, 1), 'avg_odds': round(sum(s['total_odds'] for s in slips) / total_slips, 2) if total_slips > 0 else 0, 'avg_stake': round(total_staked / total_slips, 2) if total_slips > 0 else 0, 'current_streak': current_streak, 'max_win_streak': max_win_streak, 'max_loss_streak': max_loss_streak, 'by_league': league_stats, 'best_days': [{'date': s['date'], 'profit': s['profit'], 'odds': s['total_odds']} for s in best_days], 'worst_days': [{'date': s['date'], 'profit': s['profit'], 'odds': s['total_odds']} for s in worst_days[-5:][::-1]], 'monthly': monthly_stats } def generate_single_bets(results, min_confidence=60): """ Genera scommesse singole (1 match per scommessa) Mostra la vera accuratezza del modello """ bets = [] # Filtra per confidence high_conf = [m for m in results if m['confidence'] >= min_confidence] for match in high_conf: odds = calculate_odds( match['predicted'], match['home_prob'], match['draw_prob'], match['away_prob'], match['correct'] ) stake = 10.0 actual_win = stake * odds if match['correct'] else 0 profit = actual_win - stake bets.append({ 'id': f"BET-{match['date'].replace('-', '')}-{match['home_team'][:3].upper()}", 'date': match['date'], 'time': match.get('time', '15:00'), 'home_team': match['home_team'], 'away_team': match['away_team'], 'league': match['league'], 'league_name': match['league_name'], 'flag': LEAGUE_FLAGS.get(match['league'], '⚽'), 'prediction': match['predicted'], 'predicted_sign': '1' if match['predicted'] == 'HOME' else ('X' if match['predicted'] == 'DRAW' else '2'), 'actual_result': match['actual'], 'actual_sign': '1' if match['actual'] == 'HOME' else ('X' if match['actual'] == 'DRAW' else '2'), 'home_goals': match['home_goals'], 'away_goals': match['away_goals'], 'confidence': match['confidence'], 'odds': odds, 'stake': stake, 'potential_win': round(stake * odds, 2), 'actual_win': round(actual_win, 2), 'profit': round(profit, 2), 'is_correct': match['correct'] }) return bets def calculate_single_bet_stats(bets): """Statistiche per scommesse singole""" total = len(bets) won = sum(1 for b in bets if b['is_correct']) lost = total - won total_staked = sum(b['stake'] for b in bets) total_won = sum(b['actual_win'] for b in bets) total_profit = sum(b['profit'] for b in bets) win_rate = (won / total * 100) if total > 0 else 0 roi = (total_profit / total_staked * 100) if total_staked > 0 else 0 # By league by_league = defaultdict(lambda: {'won': 0, 'lost': 0, 'profit': 0, 'staked': 0}) for bet in bets: league = bet['league'] by_league[league]['staked'] += bet['stake'] by_league[league]['profit'] += bet['profit'] if bet['is_correct']: by_league[league]['won'] += 1 else: by_league[league]['lost'] += 1 league_stats = {} for league, stats in by_league.items(): total_l = stats['won'] + stats['lost'] league_stats[league] = { 'name': LEAGUE_NAMES.get(league, league), 'flag': LEAGUE_FLAGS.get(league, '⚽'), 'total': total_l, 'won': stats['won'], 'lost': stats['lost'], 'win_rate': round(stats['won'] / total_l * 100, 1) if total_l > 0 else 0, 'profit': round(stats['profit'], 2), 'roi': round(stats['profit'] / stats['staked'] * 100, 1) if stats['staked'] > 0 else 0 } # By confidence band by_confidence = {} for band in [(60, 65), (65, 70), (70, 75), (75, 100)]: band_bets = [b for b in bets if band[0] <= b['confidence'] < band[1]] if band_bets: band_won = sum(1 for b in band_bets if b['is_correct']) band_profit = sum(b['profit'] for b in band_bets) band_staked = sum(b['stake'] for b in band_bets) by_confidence[f"{band[0]}-{band[1]}"] = { 'range': f"{band[0]}%-{band[1]}%", 'total': len(band_bets), 'won': band_won, 'win_rate': round(band_won / len(band_bets) * 100, 1), 'profit': round(band_profit, 2), 'roi': round(band_profit / band_staked * 100, 1) if band_staked > 0 else 0 } # Recent performance (last 20) recent = sorted(bets, key=lambda x: x['date'], reverse=True)[:20] recent_won = sum(1 for b in recent if b['is_correct']) recent_profit = sum(b['profit'] for b in recent) return { 'total_bets': total, 'won': won, 'lost': lost, 'win_rate': round(win_rate, 1), 'total_staked': round(total_staked, 2), 'total_won': round(total_won, 2), 'total_profit': round(total_profit, 2), 'roi': round(roi, 1), 'avg_odds': round(sum(b['odds'] for b in bets) / total, 2) if total > 0 else 0, 'by_league': league_stats, 'by_confidence': by_confidence, 'recent_20': { 'won': recent_won, 'win_rate': round(recent_won / len(recent) * 100, 1) if recent else 0, 'profit': round(recent_profit, 2) } } def main(): print("=" * 70) print("BetPredictAI - Historical Betting Slips Generator") print("=" * 70) # Load verification results verification_path = os.path.join(DATA_DIR, 'verification_results.json') if not os.path.exists(verification_path): print(" ERROR: verification_results.json not found!") print(" Run train_model_v4_robust.py first.") return with open(verification_path, 'r') as f: verification = json.load(f) results = verification.get('results', []) print(f"\n Loaded {len(results)} verified matches") # Group by date matches_by_date = group_matches_by_date(results) print(f" Spanning {len(matches_by_date)} unique dates") # ===== SINGLE BETS (recommended) ===== print("\n" + "=" * 50) print(" SINGLE BETS (Recommended Strategy)") print("=" * 50) single_bets = generate_single_bets(results, min_confidence=60) single_stats = calculate_single_bet_stats(single_bets) print(f"\n Total Single Bets: {single_stats['total_bets']}") print(f" Win Rate: {single_stats['win_rate']:.1f}%") print(f" ROI: {single_stats['roi']:+.1f}%") print(f" Total Profit: €{single_stats['total_profit']:+.2f}") single_output = { 'generated_at': datetime.now().isoformat(), 'strategy': 'single_bets', 'confidence_threshold': 60, 'statistics': single_stats, 'bets': single_bets } # ===== MULTIPLE BETS ===== print("\n" + "=" * 50) print(" MULTIPLE BETS (Higher Risk/Reward)") print("=" * 50) # Generate slips at different confidence levels confidence_levels = [50, 55, 60, 65] all_outputs = {'single_bets': single_output} for conf in confidence_levels: print(f"\n Generating slips with confidence >= {conf}%...") slips = generate_daily_slips(matches_by_date, min_confidence=conf, max_selections=4) stats = calculate_statistics(slips) output = { 'generated_at': datetime.now().isoformat(), 'confidence_threshold': conf, 'statistics': stats, 'slips': slips } all_outputs[f"conf_{conf}"] = output print(f" Generated {len(slips)} slips") print(f" Win Rate: {stats['win_rate']:.1f}%") print(f" ROI: {stats['roi']:+.1f}%") print(f" Total Profit: €{stats['total_profit']:+.2f}") # Choose best performing strategy (default 60%) best_conf = 60 primary_output = all_outputs[f"conf_{best_conf}"] # Save outputs print("\n Saving outputs...") # Save single bets (recommended strategy) with open(os.path.join(DATA_DIR, 'historical_single_bets.json'), 'w') as f: json.dump(single_output, f, indent=2) # Save primary (conf_60) as main file with open(os.path.join(DATA_DIR, 'historical_slips.json'), 'w') as f: json.dump(primary_output, f, indent=2) # Save all strategies with open(os.path.join(DATA_DIR, 'historical_slips_all.json'), 'w') as f: json.dump(all_outputs, f, indent=2) # Create a summary for the frontend frontend_stats = { 'generated_at': datetime.now().isoformat(), 'single_bets': { 'total': single_stats['total_bets'], 'won': single_stats['won'], 'win_rate': single_stats['win_rate'], 'roi': single_stats['roi'], 'profit': single_stats['total_profit'], 'by_confidence': single_stats['by_confidence'], 'by_league': single_stats['by_league'] }, 'model_accuracy': { 'overall': verification.get('summary', {}).get('accuracy', 53.8), 'by_confidence': verification.get('by_confidence', {}), 'training_matches': verification.get('training_info', {}).get('training_matches', 17611), 'test_matches': verification.get('training_info', {}).get('test_matches', 400) }, 'improvement_over_random': { 'random_probability': 33.3, # 1/3 for 3 outcomes 'ai_probability': single_stats['win_rate'], 'improvement_factor': round(single_stats['win_rate'] / 33.3, 2), 'high_confidence_accuracy': single_stats['by_confidence'].get('65-70', {}).get('win_rate', single_stats['win_rate']) } } with open(os.path.join(DATA_DIR, 'frontend_stats.json'), 'w') as f: json.dump(frontend_stats, f, indent=2) # Print summary print("\n" + "=" * 70) print("GENERATION COMPLETE") print("=" * 70) print(f"\n SINGLE BETS (Recommended):") print(f" Total Bets: {single_stats['total_bets']}") print(f" Win Rate: {single_stats['win_rate']:.1f}%") print(f" ROI: {single_stats['roi']:+.1f}%") print(f" Profit: €{single_stats['total_profit']:+.2f}") print(f"\n By Confidence Band:") for band, bs in single_stats['by_confidence'].items(): print(f" {bs['range']}: {bs['win_rate']:.1f}% ({bs['won']}/{bs['total']}) ROI: {bs['roi']:+.1f}%") print(f"\n IMPROVEMENT OVER RANDOM:") print(f" Random (1/3): 33.3%") print(f" AI System: {single_stats['win_rate']:.1f}%") print(f" Improvement: {single_stats['win_rate'] / 33.3:.2f}x") stats = primary_output['statistics'] print(f"\n MULTIPLE BETS (Higher Risk):") print(f" Total Slips: {stats['total_slips']}") print(f" Won: {stats['won_slips']} ({stats['win_rate']:.1f}%)") print(f" Lost: {stats['lost_slips']}") print(f"\n By League (Single Bets):") for league, ls in single_stats['by_league'].items(): print(f" {ls['flag']} {ls['name']}: {ls['win_rate']:.1f}% ({ls['won']}/{ls['total']})") print(f"\n Files saved:") print(f" - {os.path.join(DATA_DIR, 'historical_single_bets.json')}") print(f" - {os.path.join(DATA_DIR, 'historical_slips.json')}") print(f" - {os.path.join(DATA_DIR, 'historical_slips_all.json')}") print(f" - {os.path.join(DATA_DIR, 'frontend_stats.json')}") if __name__ == '__main__': main()