Working backtester with comparison to equally-weighted portfolio.

Author: Sinbad-The-Sailor

backtest.py

@@ -1,7 +1,11 @@
 # -*- coding: utf-8 -*-
 import time
+
+import numpy as np
 import pandas as pd
 
+from matplotlib import pyplot as plt
+
 from tests.test_config import TEST_YAHOO_STOCK_UNIVERSE_16, TEST_YAHOO_STOCK_UNIVERSE_8, TEST_YAHOO_STOCK_UNIVERSE_4
 from src.abacus.utils.portfolio import Portfolio
 from src.abacus.utils.universe import Universe
@@ -12,7 +16,8 @@
 # Backtesting setup.
 instrument_specification = {}
 inital_weights = {}
-wealth = []
+wealth = [100]
+wealth_n = [100]
 number_of_start_assets = 5
 for i, ticker in enumerate(sorted(TEST_YAHOO_STOCK_UNIVERSE_8)):
     file = f"tests/data/{ticker}.csv"
@@ -24,7 +29,7 @@
 
 # Date range for backtesting.
 start_date = "2020-01-02"
-end_date = "2020-01-05" # "2023-05-31"
+end_date = "2020-05-02" # "2023-05-31"
 date_range = pd.date_range(start=start_date, end=end_date, freq='B')
 solutions = {"2020-01-01": inital_weights}
 times = {}
@@ -34,15 +39,32 @@
     universe.date_today = date
     simulator = Simulator(universe)
     simulator.calibrate()
-    simulator.run_simulation(time_steps=10, number_of_simulations=1_000)
-    optimizer = MPCMaximumReturn(universe, portfolio, simulator.return_tensor, gamma=2, l1_penalty=0, l2_penalty=0.15,
+    simulator.run_simulation(time_steps=5, number_of_simulations=1000)
+    optimizer = MPCMaximumReturn(universe, portfolio, simulator.return_tensor, gamma=2, l1_penalty=0, l2_penalty=0.02,
                                  covariance_matrix=simulator.covariance_matrix)
     optimizer.solve()
     solution = optimizer.solution
     times[date] = time.time() - t1
     portfolio.weights = solution
     solutions[date] = solution
 
+    universe_returns = universe.todays_returns
+    portfolio_weights = np.array(list(solution.values()))
+    equal_weights = (1/8) * np.ones(8)
+    portfolio_return = np.dot(universe_returns, portfolio_weights)
+    equal_return = np.dot(universe_returns, equal_weights)
+    wealth.append(wealth[-1] * (1 + portfolio_return))
+    wealth_n.append(wealth_n[-1] * (1 + equal_return))
+
+dates = ["2024-01-01"]
+for date in date_range:
+    dates.append(str(date)[0:10])
+
+np.savetxt('data.csv', np.array(wealth), delimiter=',')
+
+plt.plot(dates, wealth)
+plt.plot(dates, wealth_n)
+plt.show()
 
 print('\n' * 10)
 for date, solution in solutions.items():

src/abacus/legacy/leg_main.py

@@ -8,9 +8,9 @@
 from utils.stock_factory import StockFactory
 from utils.data_handler import YahooDataHandler
 from utils.portfolio import Portfolio
-from utils.risk_assessor import RiskAssessor
+from abacus.assessor.risk_assessor import RiskAssessor
 from simulator.simulator import Simulator
-from optimizer.optimizer import Optimizer
+from abacus.optimizer.leg_optimizer import Optimizer
 from optimizer.enums import OptimizationModels
 
 

src/abacus/optimizer/leg_optimizer.py

@@ -56,10 +56,7 @@ def _set_ampl_data(self):
             tensor_size = price_tensor.shape
             number_of_assets = tensor_size[0]
             number_of_scenarios = tensor_size[1]
-
-            price_dict = {(j+1, asset.identifier): price_tensor[asset.id][j] for asset in assets
-                                                                             for j in range(number_of_scenarios)}
-
+            price_dict = {(j+1, asset.identifier): price_tensor[asset.id][j] for asset in assets for j in range(number_of_scenarios)}
             self._ampl.get_set("assets").set_values(asset_identifiers)
             self._ampl.param["gamma"] = -24
             self._ampl.param["risk_free_rate"] = 0.04

src/abacus/utils/instrument.py

@@ -26,7 +26,7 @@ def log_returns(self) -> pd.DataFrame:
 
     @property
     def art_returns(self) -> pd.DataFrame:
-        return self.price_history.pct_change()[1:]
+        return np.array(self.price_history.pct_change())[1:]
 
     @property
     def log_returns_tensor(self) -> torch.Tensor:

src/abacus/utils/universe.py

@@ -1,4 +1,5 @@
 # -*- coding: utf-8 -*-
+import numpy as np
 
 from pandas import DataFrame
 from datetime import date
@@ -38,6 +39,10 @@ def instruments(self) -> list[Instrument]:
             built_instruments.append(ins)
         return built_instruments
 
+    @property
+    def todays_returns(self):
+        return np.array([instrument.art_returns[-1] for instrument in self.instruments]).flatten()
+
     def has_updated_cache(self) -> bool:
         has_cache = self._instruments is not None
         has_last_build = self._instrument_build_date is not None