Analyser la combinaison de deux stratégies alpha complementaires via le QC Algorithm Framework: - EMA-Cross Alpha: Mean-reversion rapide sur 5 tech stocks (EMA20 > EMA50) - TrendStocks Alpha: Trend-following a double confirmation sur 15 actions diversifiees (Prix > SMA200 AND EMA20 > EMA50)
# Setup QuantBook
from AlgorithmImports import *
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import warnings
warnings.filterwarnings('ignore')
plt.style.use('seaborn-v0_8-darkgrid')
plt.rcParams['figure.figsize'] = (14, 5)
qb = QuantBook()
print("QuantBook initialise.")QuantBook initialise.On charge les 15 actions de l’univers combine (5 tech + 10 diversifiees).
# Univers combine: 5 tech stocks + 10 diversifiees
ema_tickers = ["AAPL", "MSFT", "GOOGL", "AMZN", "NVDA"]
trend_tickers = [
"AAPL", "MSFT", "GOOGL", "AMZN", "NVDA", # Tech (overlap)
"JPM", "V", "MA", # Financials
"UNH", "JNJ", # Healthcare
"XOM", "CVX", # Energy
"HD", "PG", "KO" # Consumer staples
]
# Benchmark pour la comparaison
all_tickers = list(set(trend_tickers)) + ["SPY"]
symbols = {}
for ticker in all_tickers:
symbols[ticker] = qb.add_equity(ticker, Resolution.DAILY).symbol
# Charger l'historique (2015-2026)
start = datetime(2015, 1, 1)
end = datetime(2026, 1, 1)
history = qb.history(list(symbols.values()), start, end, Resolution.DAILY)
print(f"Donnees chargees: {len(history)} lignes")Donnees chargees: 8298 lignesPivot de la série ‘close’ en DataFrame large, avec remapping des colonnes Symbol → ticker pour Framework_Composite_EMATrend.
# Pivoter les donnees pour avoir un DataFrame de prix
closes = history['close'].unstack(level=0)
# Renommer les colonnes
symbol_to_ticker = {str(v): k for k, v in symbols.items()}
closes.columns = [symbol_to_ticker.get(str(c), str(c)) for c in closes.columns]
closes = closes.dropna()
print(f"Periode: {closes.index[0].date()} a {closes.index[-1].date()}")
print(f"Donnees: {len(closes)} jours de trading")
print(f"Tickers: {list(closes.columns)}")Periode: 2015-01-02 a 2025-12-31
Donnees: 2766 jours de trading
Tickers: ['AAPL', 'GOOGL', 'SPY']On implemente la logique des deux AlphaModels dans un format vectorise pour le backtest.
def compute_ema_cross_signals(closes, tickers, fast=20, slow=50):
"""Genere les signaux EMA-Cross (daily)."""
signals = pd.DataFrame(index=closes.index, columns=tickers)
for ticker in tickers:
if ticker not in closes.columns:
continue
ema_fast = closes[ticker].ewm(span=fast, adjust=False).mean()
ema_slow = closes[ticker].ewm(span=slow, adjust=False).mean()
signals[ticker] = (ema_fast > ema_slow).astype(int)
return signals
def compute_trend_stocks_signals(closes, tickers, ema_fast=20, ema_slow=50, sma_trend=200):
"""Genere les signaux TrendStocks (double confirmation)."""
signals = pd.DataFrame(index=closes.index, columns=tickers)
for ticker in tickers:
if ticker not in closes.columns:
continue
price = closes[ticker]
ema_f = price.ewm(span=ema_fast, adjust=False).mean()
ema_s = price.ewm(span=ema_slow, adjust=False).mean()
sma_200 = price.rolling(sma_trend).mean()
price_above_sma = price > sma_200
ema_bullish = ema_f > ema_s
signals[ticker] = (price_above_sma & ema_bullish).astype(int)
return signals
# Generer les signaux
ema_signals = compute_ema_cross_signals(closes, ema_tickers)
trend_signals = compute_trend_stocks_signals(closes, trend_tickers)
print("Signaux EMA-Cross (derniers 5 jours):")
print(ema_signals.iloc[-5:])
print(f"\nSignaux TrendStocks (derniers 5 jours):")
print(trend_signals.iloc[-5:])Signaux EMA-Cross (derniers 5 jours):
AAPL MSFT GOOGL AMZN NVDA
time
2025-12-24 13:00:00 0 NaN 1 NaN NaN
2025-12-26 16:00:00 0 NaN 1 NaN NaN
2025-12-29 16:00:00 0 NaN 1 NaN NaN
2025-12-30 16:00:00 0 NaN 1 NaN NaN
2025-12-31 16:00:00 0 NaN 1 NaN NaN
Signaux TrendStocks (derniers 5 jours):
AAPL MSFT GOOGL AMZN NVDA JPM V MA UNH JNJ XOM \
time
2025-12-24 13:00:00 0 NaN 0 NaN NaN NaN NaN NaN NaN NaN NaN
2025-12-26 16:00:00 0 NaN 0 NaN NaN NaN NaN NaN NaN NaN NaN
2025-12-29 16:00:00 0 NaN 0 NaN NaN NaN NaN NaN NaN NaN NaN
2025-12-30 16:00:00 0 NaN 0 NaN NaN NaN NaN NaN NaN NaN NaN
2025-12-31 16:00:00 0 NaN 0 NaN NaN NaN NaN NaN NaN NaN NaN
CVX HD PG KO
time
2025-12-24 13:00:00 NaN NaN NaN NaN
2025-12-26 16:00:00 NaN NaN NaN NaN
2025-12-29 16:00:00 NaN NaN NaN NaN
2025-12-30 16:00:00 NaN NaN NaN NaN
2025-12-31 16:00:00 NaN NaN NaN NaN Les 5 tech stocks (AAPL, MSFT, GOOGL, AMZN, NVDA) apparaissent dans les deux signaux, creant une synergie potentielle.
Fonction de backtest vectorise combinant les deux stratégies avec allocation variable.
def backtest_composite(closes, ema_signals, trend_signals,
ema_alloc=0.40, trend_alloc=0.60,
ema_rebal_freq=1, trend_rebal_freq=5):
"""
Backtest du composite avec allocation variable.
Args:
ema_alloc: Allocation a EMA-Cross (ex: 0.40 = 40%)
trend_alloc: Allocation a TrendStocks (ex: 0.60 = 60%)
ema_rebal_freq: Frequence rebalance EMA (1 = daily)
trend_rebal_freq: Frequence rebalance Trend (5 = weekly)
"""
returns_df = closes.pct_change()
portfolio_values = [1.0]
warmup = 200
ema_counter = 0
trend_counter = 0
# Tracking des positions
ema_holdings = set()
trend_holdings = set()
for i in range(warmup, len(closes)):
# Mise a jour EMA-Cross holdings (daily)
ema_counter += 1
if ema_counter >= ema_rebal_freq:
ema_counter = 0
ema_holdings = set()
for t in ema_tickers:
if t in ema_signals.columns and ema_signals[t].iloc[i] == 1:
ema_holdings.add(t)
# Mise a jour TrendStocks holdings (weekly)
trend_counter += 1
if trend_counter >= trend_rebal_freq:
trend_counter = 0
trend_holdings = set()
for t in trend_tickers:
if t in trend_signals.columns and trend_signals[t].iloc[i] == 1:
trend_holdings.add(t)
# Calcul du return du portefeuille
port_return = 0.0
# EMA-Cross contribution
if len(ema_holdings) > 0:
weight_per_stock = ema_alloc / len(ema_holdings)
for t in ema_holdings:
if t in returns_df.columns:
port_return += weight_per_stock * returns_df[t].iloc[i]
# TrendStocks contribution
if len(trend_holdings) > 0:
weight_per_stock = trend_alloc / len(trend_holdings)
for t in trend_holdings:
if t in returns_df.columns:
port_return += weight_per_stock * returns_df[t].iloc[i]
portfolio_values.append(portfolio_values[-1] * (1 + port_return))
# Calcul des metriques
returns = np.diff(portfolio_values) / np.array(portfolio_values[:-1])
cum_returns = pd.Series(portfolio_values[1:], index=closes.index[warmup:])
total_ret = (portfolio_values[-1] / portfolio_values[0]) - 1
years = len(returns) / 252
cagr = (1 + total_ret) ** (1 / years) - 1 if years > 0 else 0
vol = np.std(returns) * np.sqrt(252) if len(returns) > 1 else 0
sharpe = (cagr - 0.03) / vol if vol > 0.001 else 0
running_max = cum_returns.expanding().max()
drawdown = (cum_returns - running_max) / running_max
max_dd = drawdown.min()
return {
'cum': cum_returns,
'sharpe': sharpe,
'cagr': cagr,
'max_dd': max_dd,
'vol': vol,
'final_value': portfolio_values[-1]
}
print("Fonction de backtest composite definie.")Fonction de backtest composite definie.On teste différentes allocations EMA/Trend de 30/70 a 70/30.
# Test differentes allocations
allocations = [
(0.30, 0.70, "EMA30/Trend70"),
(0.40, 0.60, "EMA40/Trend60"),
(0.50, 0.50, "EMA50/Trend50"),
(0.60, 0.40, "EMA60/Trend40"),
(0.70, 0.30, "EMA70/Trend30"),
]
results = {}
print(f"{'Allocation':<20} {'Sharpe':>8} {'CAGR':>8} {'MaxDD':>8} {'Vol':>8}")
print("-" * 55)
for ema_alloc, trend_alloc, name in allocations:
r = backtest_composite(closes, ema_signals, trend_signals,
ema_alloc=ema_alloc, trend_alloc=trend_alloc)
results[name] = r
print(f"{name:<20} {r['sharpe']:>8.3f} {r['cagr']:>7.1%} {r['max_dd']:>7.1%} {r['vol']:>7.1%}")
# Trouver la meilleure allocation
best_alloc = max(results.items(), key=lambda x: x[1]['sharpe'])
print(f"\nMeilleure allocation: {best_alloc[0]} (Sharpe={best_alloc[1]['sharpe']:.3f})")Allocation Sharpe CAGR MaxDD Vol
-------------------------------------------------------
EMA30/Trend70 0.424 9.3% -20.9% 14.9%
EMA40/Trend60 0.444 9.6% -21.0% 14.9%
EMA50/Trend50 0.462 9.9% -21.2% 15.0%
EMA60/Trend40 0.480 10.2% -21.4% 15.0%
EMA70/Trend30 0.497 10.5% -21.7% 15.0%
Meilleure allocation: EMA70/Trend30 (Sharpe=0.497)L’allocation optimale depend de la correlation entre les deux stratégies.
Verdict: L’allocation avec le Sharpe le plus eleve est le meilleur compromis.
def backtest_single_strategy(closes, signals, alloc=1.0, rebal_freq=1):
"""Backtest d'une seule strategy."""
returns_df = closes.pct_change()
portfolio_values = [1.0]
warmup = 200
counter = 0
holdings = set()
for i in range(warmup, len(closes)):
counter += 1
if counter >= rebal_freq:
counter = 0
holdings = set()
for t in signals.columns:
if signals[t].iloc[i] == 1:
holdings.add(t)
port_return = 0.0
if len(holdings) > 0:
weight_per_stock = alloc / len(holdings)
for t in holdings:
if t in returns_df.columns:
port_return += weight_per_stock * returns_df[t].iloc[i]
portfolio_values.append(portfolio_values[-1] * (1 + port_return))
returns = np.diff(portfolio_values) / np.array(portfolio_values[:-1])
total_ret = (portfolio_values[-1] / portfolio_values[0]) - 1
years = len(returns) / 252
cagr = (1 + total_ret) ** (1 / years) - 1 if years > 0 else 0
vol = np.std(returns) * np.sqrt(252)
sharpe = (cagr - 0.03) / vol if vol > 0.001 else 0
return {'sharpe': sharpe, 'cagr': cagr, 'values': portfolio_values}
# Backtester les strategies individuelles
ema_result = backtest_single_strategy(closes, ema_signals, alloc=1.0, rebal_freq=1)
trend_result = backtest_single_strategy(closes, trend_signals, alloc=1.0, rebal_freq=5)
print("Strategies individuelles:")
print(f" EMA-Cross: Sharpe={ema_result['sharpe']:.3f}, CAGR={ema_result['cagr']:.1%}")
print(f" TrendStocks: Sharpe={trend_result['sharpe']:.3f}, CAGR={trend_result['cagr']:.1%}")
print(f"\nComposite ({best_alloc[0]}): Sharpe={best_alloc[1]['sharpe']:.3f}, CAGR={best_alloc[1]['cagr']:.1%}")Strategies individuelles:
EMA-Cross: Sharpe=0.543, CAGR=11.3%
TrendStocks: Sharpe=0.362, CAGR=8.5%
Composite (EMA70/Trend30): Sharpe=0.497, CAGR=10.5%fig, axes = plt.subplots(1, 2, figsize=(16, 5))
# Gauche: Comparaison des allocations
ax = axes[0]
for name, r in results.items():
ax.plot(r['cum'].values, label=f"{name} (S={r['sharpe']:.2f})", linewidth=1.5)
ax.set_title('Allocation EMA/Trend', fontsize=12, fontweight='bold')
ax.set_ylabel('Valeur du portefeuille')
ax.legend(fontsize=9)
ax.grid(True, alpha=0.3)
# Droite: Composite vs Strategies individuelles
ax = axes[1]
ax.plot(ema_result['values'][200:], label=f"EMA-Cross (S={ema_result['sharpe']:.2f})", linewidth=1.5)
ax.plot(trend_result['values'][200:], label=f"TrendStocks (S={trend_result['sharpe']:.2f})", linewidth=1.5)
ax.plot(best_alloc[1]['cum'].values, label=f"Composite {best_alloc[0]} (S={best_alloc[1]['sharpe']:.2f})",
linewidth=2, linestyle='--')
ax.set_title('Synergie Composite', fontsize=12, fontweight='bold')
ax.set_ylabel('Valeur du portefeuille')
ax.legend(fontsize=9)
ax.grid(True, alpha=0.3)
plt.tight_layout()
plt.savefig('composite_ematrend_analysis.png', dpi=150, bbox_inches='tight')
plt.show()
print("Graphique sauvegarde.")
Graphique sauvegarde.# Analyser la correlation des signaux sur les tech stocks (overlap)
tech_overlap = ["AAPL", "MSFT", "GOOGL", "AMZN", "NVDA"]
# Compter les accords entre les deux strategies
agreement = pd.DataFrame(index=ema_signals.index, columns=tech_overlap)
for t in tech_overlap:
if t in ema_signals.columns and t in trend_signals.columns:
agreement[t] = (ema_signals[t] == trend_signals[t]) & (ema_signals[t] == 1)
# Pourcentage d'accord par stock
agreement_pct = agreement.mean() * 100
print("Accord EMA-Cross / TrendStocks sur les tech stocks:")
print(f"{'Stock':<10} {'Accord (%)':>12}")
print("-" * 25)
for t in tech_overlap:
if t in agreement_pct.index:
print(f"{t:<10} {agreement_pct[t]:>11.1f}%")
avg_agreement = agreement_pct.mean()
print(f"\nAccord moyen: {avg_agreement:.1f}%")
print(f"\nInterpretation: Quand les deux strategies sont d'accord sur un tech stock,\n"
f"ce stock recoit une allocation double dans le composite.")Accord EMA-Cross / TrendStocks sur les tech stocks:
Stock Accord (%)
-------------------------
AAPL 34.2%
MSFT 0.0%
GOOGL 42.3%
AMZN 0.0%
NVDA 0.0%
Accord moyen: 15.3%
Interpretation: Quand les deux strategies sont d'accord sur un tech stock,
ce stock recoit une allocation double dans le composite.| Metrique | EMA-Cross | TrendStocks | Composite optimal |
|---|---|---|---|
| Sharpe | (a remplir) | (a remplir) | (a remplir) |
| CAGR | (a remplir) | (a remplir) | (a remplir) |
| Max DD | (a remplir) | (a remplir) | (a remplir) |
Allocation: [a remplir avec la meilleure allocation]