try:
# Black Litterman output (HMM views included)
r_pred = ss.get('returns', 'predicted')
covariance_pred = ss.get('covariance', 'predicted')
volumes_pred = ss.get('volumes', 'predicted')
sigmas_pred = ss.get('sigmas', 'predicted')
# Predicted costs
optimization_tcost = cp.TcostModel(half_spread=0.0005 / 2.,
nonlin_coeff=1.,
sigma=sigmas_pred,
volume=volumes_pred)
optimization_hcost = cp.HcostModel(borrow_costs=0.0001)
# Covariance setup
bl_risk_model = cp.FullSigma(covariance_pred)
# Black Litterman policy
logging.warning('Running backtest')
# Optimization parameters
leverage_limit = cp.LeverageLimit(1)
fully_invested = cp.ZeroCash()
long_only = cp.LongOnly()
# Optimization policy
if mode == 'c':
scns = 1
else:
scns = 5
bl_mps_policy = cp.MultiPeriodScenarioOpt(
end_date=end_date)['Adj. Close'])
for ticker in tickers]))
returns = prices.pct_change()
returns[["USDOLLAR"]] = quandl.get(
'FRED/DTB3', start_date=start_date, end_date=end_date) / (250 * 100)
returns = returns.fillna(method='ffill').iloc[1:]
returns = returns[-251:]
r_hat = returns.rolling(window=250, min_periods=250).mean().shift(1).dropna()
Sigma_hat = returns.rolling(window=250, min_periods=250).cov().dropna()
tcost_model = cp.TcostModel(half_spread=10E-4)
hcost_model = cp.HcostModel(borrow_costs=1E-4)
risk_model = cp.FullSigma(Sigma_hat)
gamma_risk, gamma_trade, gamma_hold = 5., 1.5, 1.
leverage_limit = cp.LeverageLimit(3)
spo_policy = cp.SinglePeriodOpt(return_forecast=r_hat,
costs=[
gamma_risk * risk_model,
gamma_trade * tcost_model,
gamma_hold * hcost_model
],
constraints=[leverage_limit])
# mpo_policy = cp.MultiPeriodOpt
current_portfolio = pd.Series(index=r_hat.columns, data=0)
current_portfolio.USDOLLAR = 10000
# Equilibrium results
start_date = datetime.strptime(bl_ewm_model.cfg['start_date'], '%Y%m%d') + \
timedelta(days=bl_ewm_model.cfg['train_len'] * 1.75)
end_date = datetime.strptime(bl_ewm_model.cfg['end_date'], '%Y%m%d')
w_equal = pd.Series(index=r_pred.columns, data=[1] * len(r_pred.columns))
w_equal.loc['USDOLLAR'] = 0.
w_equal = w_equal / sum(w_equal)
optimization_tcost = cp.TcostModel(half_spread=0.0005/2., nonlin_coeff=1.,
sigma=sigmas_pred,
volume=volumes_pred)
optimization_hcost=cp.HcostModel(borrow_costs=0.0001)
if bl_ewm_model.cfg['covariance']['method'] == 'SS':
spo_risk_model = cp.FullSigma(bl_ewm_model.get('covariance', 'predicted'))
elif bl_ewm_model.cfg['covariance']['method'] == 'FF5':
spo_risk_model = cp.FactorModelSigma(bl_ewm_model.get('exposures', 'predicted'),
bl_ewm_model.get('factor_sigma', 'predicted'),
bl_ewm_model.get('idyos', 'predicted'))
else:
raise NotImplemented('The %s risk model is not implemented yet'.format(bl_ewm_model.cfg['risk']))
logging.warning('Running simulation...')
# Optimization parameters
gamma_risk, gamma_trade, gamma_hold = 5., 15., 1.
leverage_limit = cp.LeverageLimit(1)
fully_invested = cp.ZeroCash()
long_only = cp.LongOnly()
# Optimization policy
try:
r_pred = ss.get('returns', 'predicted')
volumes_pred = ss.get('volumes', 'predicted')
sigmas_pred = ss.get('sigmas', 'predicted')
# Predicted costs
optimization_tcost = cp.TcostModel(half_spread=0.0005 / 2.,
nonlin_coeff=1.,
sigma=sigmas_pred,
volume=volumes_pred)
optimization_hcost = cp.HcostModel(borrow_costs=0.0001)
# Covariance setup
if ss.cfg['covariance']['method'] == 'SS':
risk_model = cp.FullSigma(ss.get('covariance', 'predicted'))
elif ss.cfg['covariance']['method'] == 'FF5':
risk_model = cp.FactorModelSigma(
ss.get('exposures', 'predicted'),
ss.get('factor_sigma', 'predicted'),
ss.get('idyos', 'predicted'))
else:
raise NotImplemented(
'The %s risk model is not implemented yet'.format(
ss.cfg['covariance']))
# Black Litterman policy
logging.warning('Running backtest')
# Optimization parameters
leverage_limit = cp.LeverageLimit(1)