import pandas as pd import numpy as np import os seasons = ['1920','2021','2122','2223','2324','2425','2526'] league_names = { 'E0':'EPL','E1':'Championship','E2':'League One','E3':'League Two','EC':'Conference', 'SP1':'La Liga','SP2':'Segunda','I1':'Serie A','I2':'Serie B', 'D1':'Bundesliga','D2':'2.Bundesliga','F1':'Ligue 1','F2':'Ligue 2', 'N1':'Eredivisie','B1':'Belgian Pro League','P1':'Primeira Liga', 'T1':'Super Lig','G1':'Greek SL', 'SC0':'Scottish Prem','SC1':'Scottish Champ','SC2':'Scottish L1','SC3':'Scottish L2', } def simulate(bets): if not bets: return 0,0.0,0.0,0.0 profits = [(o-1) if w else -1 for o,w in bets] n = len(profits); total = sum(profits) return n, round(total,2), round(total/n*100,2), round(sum(w for _,w in bets)/n*100,1) def remove_margin(h,d,a): over = 1/h+1/d+1/a return 1/h/over, 1/d/over, 1/a/over # Load all data all_dfs = [] for code, name in league_names.items(): for s in seasons: path = f'E:/soccer/league_data/{code}_{s}.csv' try: d = pd.read_csv(path, encoding='utf-8-sig') if len(d) < 10 or 'FTR' not in d.columns or 'B365A' not in d.columns: continue d = d.dropna(subset=['FTR','B365A','PSCH','PSCD','PSCA']) if len(d) < 10: continue d['Season'] = s; d['LeagueName'] = name all_dfs.append(d) except: pass df = pd.concat(all_dfs, ignore_index=True) df['Date'] = pd.to_datetime(df['Date'], dayfirst=True, errors='coerce') df = df.dropna(subset=['Date']).sort_values('Date').reset_index(drop=True) tH,tD,tA = zip(*df.apply(lambda r: remove_margin(r['PSCH'],r['PSCD'],r['PSCA']), axis=1)) df = df.copy() df['PSC_fair_H'] = 1/pd.Series(tH,index=df.index) df['PSC_fair_D'] = 1/pd.Series(tD,index=df.index) df['PSC_fair_A'] = 1/pd.Series(tA,index=df.index) df['edge_A'] = (df['B365A']-df['PSC_fair_A'])/df['PSC_fair_A'] df['edge_H'] = (df['B365H']-df['PSC_fair_H'])/df['PSC_fair_H'] df['edge_D'] = (df['B365D']-df['PSC_fair_D'])/df['PSC_fair_D'] # Build best-outcome bets per league (same as before) best_config = { '2.Bundesliga': ('Home','edge_H','B365H','H'), 'Belgian Pro League': ('Away','edge_A','B365A','A'), 'Bundesliga': ('Home','edge_H','B365H','H'), 'Championship': ('Draw','edge_D','B365D','D'), 'Conference': ('Home','edge_H','B365H','H'), 'EPL': ('Away','edge_A','B365A','A'), 'Eredivisie': ('Home','edge_H','B365H','H'), 'Greek SL': ('Draw','edge_D','B365D','D'), 'La Liga': ('Home','edge_H','B365H','H'), 'League One': ('Home','edge_H','B365H','H'), 'League Two': ('Home','edge_H','B365H','H'), 'Ligue 1': ('Away','edge_A','B365A','A'), 'Ligue 2': ('Home','edge_H','B365H','H'), 'Primeira Liga': ('Away','edge_A','B365A','A'), 'Scottish Champ': ('Draw','edge_D','B365D','D'), 'Scottish L1': ('Away','edge_A','B365A','A'), 'Scottish L2': ('Draw','edge_D','B365D','D'), 'Scottish Prem': ('Home','edge_H','B365H','H'), 'Segunda': ('Home','edge_H','B365H','H'), 'Serie A': ('Draw','edge_D','B365D','D'), 'Serie B': ('Home','edge_H','B365H','H'), 'Super Lig': ('Draw','edge_D','B365D','D'), } rows = [] for name, (outcome, edge_col, odds_col, result) in best_config.items(): sub = df[(df['LeagueName']==name) & (df[edge_col]>0.05)] for _, r in sub.iterrows(): tH2,tD2,tA2 = remove_margin(r['PSCH'],r['PSCD'],r['PSCA']) p = {'H':tH2,'D':tD2,'A':tA2}[result] b = r[odds_col]-1 edge = r[odds_col]*p-1 kelly_f = edge/b if b>0 else 0 if kelly_f > 0: rows.append({'Date':r['Date'],'Season':r['Season'],'League':name, 'Outcome':outcome,'Odds':r[odds_col],'Kelly':kelly_f, 'Won':r['FTR']==result}) bets_df = pd.DataFrame(rows).sort_values('Date').reset_index(drop=True) bets_df['Week'] = bets_df['Date'].dt.to_period('W') # --------------------------------------------------------------- # SEASON-BY-SEASON: continuous bankroll # --------------------------------------------------------------- KELLY_MULT = 0.25 bankroll = 1000.0 season_snapshots = {} season_start_br = {} week_history = [] for week, grp in bets_df.groupby('Week'): season = grp['Season'].mode()[0] if season not in season_start_br: season_start_br[season] = bankroll wpl = 0; wins = 0; staked = 0 for _, bet in grp.iterrows(): f = min(bet['Kelly']*KELLY_MULT, 0.25) stake = bankroll*f staked += stake if bet['Won']: wpl += stake*(bet['Odds']-1); wins += 1 else: wpl -= stake bankroll += wpl week_history.append({'Week':str(week),'Season':season,'N':len(grp), 'Wins':wins,'PL':wpl,'Bankroll':bankroll}) # End-of-season bankroll = last week of each season hist_df = pd.DataFrame(week_history) for s in seasons: sub = hist_df[hist_df['Season']==s] if not sub.empty: season_snapshots[s] = sub['Bankroll'].iloc[-1] # --------------------------------------------------------------- # PRINT: season by season # --------------------------------------------------------------- print("="*80) print("SEASON-BY-SEASON RETURNS — Best CLV outcome per league, 0.25x Kelly") print("Continuous bankroll (not reset each season)") print("="*80) print(f"\n{'Season':<10} {'Start £':>12} {'End £':>14} {'Return':>8} {'Bets':>6} {'WR%':>6} {'Max DD%':>9}") print("-"*80) prev_br = 1000.0 for s in seasons: sub_bets = bets_df[bets_df['Season']==s] sub_hist = hist_df[hist_df['Season']==s] if sub_hist.empty: continue start_br = season_start_br.get(s, prev_br) end_br = season_snapshots.get(s, prev_br) ret_pct = (end_br/start_br - 1)*100 n_bets = len(sub_bets) wr = sub_bets['Won'].mean()*100 if n_bets>0 else 0 # Max drawdown within season br_series = sub_hist['Bankroll'] peak = br_series.cummax() dd = ((peak - br_series)/peak*100).max() marker = " +" if ret_pct > 0 else " --" print(f"{s:<10} £{start_br:>11,.2f} £{end_br:>13,.2f} {ret_pct:>+7.1f}%{marker} {n_bets:>5} {wr:>5.1f}% {dd:>7.1f}%") prev_br = end_br print(f"\n{'TOTAL':<10} £{'1,000.00':>11} £{prev_br:>13,.2f} {(prev_br/1000-1)*100:>+7.1f}%") # --------------------------------------------------------------- # ALSO: flat £1/unit per bet for comparison (no compounding) # --------------------------------------------------------------- print("\n" + "="*80) print("FLAT BETTING (£1/unit per bet) — season by season, no compounding") print("="*80) print(f"\n{'Season':<10} {'Bets':>5} {'P&L':>8} {'ROI%':>7} {'WR%':>6}") print("-"*80) total_bets=[]; total_pl=0 for s in seasons: sub = bets_df[bets_df['Season']==s] bets = [(r['Odds'], r['Won']) for _,r in sub.iterrows()] n,p,roi,wr = simulate(bets) marker = " +" if roi>0 else "" print(f"{s:<10} {n:>5} {p:>8.2f} {roi:>7.2f}% {wr:>5.1f}%{marker}") total_pl += p; total_bets.extend(bets) n_all,p_all,roi_all,wr_all = simulate(total_bets) print(f"{'ALL':<10} {n_all:>5} {p_all:>8.2f} {roi_all:>7.2f}% {wr_all:>5.1f}%") # --------------------------------------------------------------- # PER-LEAGUE per-season heatmap (flat ROI) # --------------------------------------------------------------- print("\n" + "="*95) print("PER-LEAGUE PER-SEASON ROI% (flat betting, best outcome per league, >5% edge)") print("="*95) header = f"{'League':<22}" + "".join(f" {s:>8}" for s in seasons) + f" {'ALL':>8}" print(header) print("-"*95) for name, (outcome, edge_col, odds_col, result) in sorted(best_config.items()): row_str = f"{name:<22}" all_bets = [] for s in seasons: sub = bets_df[(bets_df['League']==name)&(bets_df['Season']==s)] bets = [(r['Odds'],r['Won']) for _,r in sub.iterrows()] n,p,roi,wr = simulate(bets) all_bets.extend(bets) if n >= 5: row_str += f" {roi:>+7.1f}%" else: row_str += f" {'n/a':>8}" n,p,roi,wr = simulate(all_bets) row_str += f" {roi:>+7.1f}%" print(row_str)