Overall Statistics |
Total Trades 22073 Average Win 0.09% Average Loss -0.08% Compounding Annual Return 23.030% Drawdown 67.300% Expectancy 0.117 Net Profit 181.877% Sharpe Ratio 0.763 Probabilistic Sharpe Ratio 21.765% Loss Rate 48% Win Rate 52% Profit-Loss Ratio 1.13 Alpha 0.27 Beta -0.101 Annual Standard Deviation 0.333 Annual Variance 0.111 Information Ratio 0.263 Tracking Error 0.382 Treynor Ratio -2.522 Total Fees $366576.28 |
# # Original File: # QUANTCONNECT.COM - Democratizing Finance, Empowering Individuals. # Lean Algorithmic Trading Engine v2.0. Copyright 2014 QuantConnect # Corporation. # # Changes: # The universe selection model is extended to take parameters as # optional arguments. # Ostirion.net Copyright 2021 # Hector Barrio - hbarrio@ostirion.net. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at: # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. # from QuantConnect.Data.UniverseSelection import * from Selection.FundamentalUniverseSelectionModel import FundamentalUniverseSelectionModel from itertools import groupby from math import ceil from clr import AddReference import numpy as np from typing import List, Set, Tuple, Dict AddReference("System") AddReference("QuantConnect.Common") AddReference("QuantConnect.Algorithm.Framework") class FlexibleUniverseSelectionModel(FundamentalUniverseSelectionModel): ''' Class representing a parametrically selected securities universe. Attributes: n_coarse (int): Number of securities in the coarse selection. n_fine (int): Number of securities in fine selection. age (int): Minimum time since IPO. recent (int): Maximum time from IPO. vol_lim (float): Minimum daily volume of each security. min_price (float): Minimum price of each security. max_price (float): Maximum price of each security. period (str): "Month" or "Day". Recalculate the universe every period. m_cap_lim (float): Minimum market cap of security to be considered. markets (list[str]): Markets in which the security trades. c_id (str): Code of the country of origin of securities. from_top (bool): Take the top (True) or bottom (False) volume securities. restrict_country (bool): Restrict the country of origin and market for securities. verbose (bool): False for silent, True for announcing size and components. ''' def __init__(self: None, n_coarse: int=1000, n_fine: int=500, age: int=1250, recent: int=-1, vol_lim: int=0, min_price: int=0, max_price: float=np.Inf, period: str='Month', m_cap_lim: float=5e8, markets: List[str]=["NYS", "NAS"], c_id: str='USA', from_top: bool=True, restrict_country: bool=True, verbose: bool=False, filterFineData: bool=True, universeSettings: UniverseSettings=None, securityInitializer: SecurityInitializer=None) -> None: super().__init__(filterFineData, universeSettings, securityInitializer) # Parameter settings: self.n_symbols_coarse = n_coarse self.n_symbols_fine = n_fine self.age = age self.recent = recent self.vol_lim = vol_lim self.min_price = min_price self.max_price = max_price self.period = period self.m_cap_lim = m_cap_lim self.markets = markets self.c_id = c_id self.reverse = from_top self.restrict_country = restrict_country self.verbose = verbose self.usd_vol = {} self.last_month = -1 def SelectCoarse(self, algorithm: QCAlgorithm, coarse: CoarseFundamental) -> FineFundamental: ''' Coarse unviverse selection method. Args: algorithm (QCAlgorithm): Current algorithm instance. coarse (CoarseFundamental): QC Coarse universe object. Returns: fine (FineFundamental): QC fine universe object. ''' if self.period == 'Month': if algorithm.Time.month == self.last_month: return Universe.Unchanged elif self.period != 'Day': algoithm.Log('Period not valid.. Choose "Day" or "Month". Defaulting to "Month".') c = coarse usd_vol = sorted([x for x in c if x.HasFundamentalData and x.Volume > self.vol_lim and self.max_price > x.Price > self.min_price], key=lambda x: x.DollarVolume, reverse=self.reverse)[:self.n_symbols_coarse] self.usd_vol = {x.Symbol: x.DollarVolume for x in usd_vol} if len(self.usd_vol) == 0: return Universe.Unchanged return list(self.usd_vol.keys()) def SelectFine(self, algorithm: QCAlgorithm, fine: FineFundamental) -> FineFundamental: ''' Coarse unviverse selection method. Args: algorithm (QCAlgorithm): Current algorithm instance. fine (FineFundamental): QC fine universe object. Returns: new_universe (FineFundamental): QC fine universe object. ''' f = fine a = algorithm sort_sector = sorted([x for x in f if x.MarketCap > self.m_cap_lim], key=lambda x: x.CompanyReference.IndustryTemplateCode) count = len(sort_sector) if count == 0: return Universe.Unchanged if self.recent != -1: sort_sector = [x for x in sort_sector if (a.Time - x.SecurityReference.IPODate).days < self.recent] else: sort_sector = [x for x in sort_sector if (a.Time - x.SecurityReference.IPODate).days > self.age] if self.restrict_country: sort_sector = [x for x in sort_sector if x.CompanyReference.CountryId == self.c_id and x.CompanyReference.PrimaryExchangeID in self.markets] self.last_month = a.Time.month percent = self.n_symbols_fine / count sort_usd_vol = [] for c, g in groupby(sort_sector, lambda x: x.CompanyReference.IndustryTemplateCode): y = sorted(g, key=lambda x: self.usd_vol[x.Symbol], reverse=self.reverse) c = ceil(len(y) * percent) sort_usd_vol.extend(y[:c]) sort_usd_vol = sorted(sort_usd_vol, key=lambda x: self.usd_vol[x.Symbol], reverse=self.reverse) new_universe = [x.Symbol for x in sort_usd_vol[:self.n_symbols_fine]] if self.verbose: for s in new_universe: algorithm.Log('Adding: '+str(s.Symbol)) algorithm.Log('Universe members: ' + str(len(new_universe))) return new_universe
import time from datetime import date from FlexibleUniverseSelectionModel import FlexibleUniverseSelectionModel as fsum class MLVolatilityPredictor(QCAlgorithm): def Initialize(self): YEARS = 5 self.SetStartDate(datetime.today() - timedelta(days=YEARS*365)) self.SetEndDate(datetime.today()) self.SetCash(1000000) self.SetBrokerageModel(BrokerageName.AlphaStreams) res = Resolution.Daily self.SetBenchmark("SPY") self.AddUniverseSelection(fsum(n_fine=50, period='Month', max_price=5, )) self.UniverseSettings.Resolution = res self.AddAlpha(UniverseBalancer()) self.SetExecution(ImmediateExecutionModel()) self.SetPortfolioConstruction(InsightWeightingPortfolioConstructionModel()) class UniverseBalancer(AlphaModel): """ """ def __init__(self): self.Name = 'Universe Balancer' self.removed_symbols = False def Update(self, algorithm, data): insights = [] t_delta = timedelta(days=22) for s in data.keys(): if s in self.removed_symbols: continue if algorithm.Securities[s].Invested: continue if s.Value == 'CHK': continue insights.append(Insight(s, t_delta, InsightType.Price, InsightDirection.Up, 0.02, 1, self.Name, 1)) return insights def OnSecuritiesChanged(self, algorithm, changes): self.removed_symbols = [sec.Symbol for sec in changes.RemovedSecurities]