基于期权偏度(Options Skewness)的量化交易策略详解：捕捉市场情绪的关键｜附回测代码

#region importsfrom AlgorithmImports import *#endregion## Stocks with liquid options are used as the investment universe.# Skewness is calculated as the difference between the implied volatilities of OTM puts and ATM calls.# A put option is defined as OTM when the ratio of strike price to the stock price is closest to 0.95 (but lower than 0.95),# and a call option is defined as ATM when the ratio of strike price to the stock price is closest to 1.# Only options with a time to expiration between 10 and 60 days are included.# Weekly SKEW is calculated by averaging the daily SKEW over a week.# Every week the investors sort all sample firms into quintile portfolios based on the previous week’s average skew (Tuesday close to Tuesday close).# Portfolio 1 includes firms with the lowest skew, and portfolio 5 includes firms with the highest skew.# He/she then goes long on portfolio one and short on portfolio 5 (value-weighted portfolios) and rebalances the portfolios weekly.## QC Implementation:#   - Investment universe consists of options of top 50 liquid US stocks with price > 5$.
import numpy as np
class OptionsSkewnessPredictsConsecutiveStocksReturns(QCAlgorithm):
    def Initialize(self):        self.SetStartDate(2010, 1, 1)        self.SetCash(100000)                self.min_expiry = 40        self.max_expiry = 60                self.period = 5             # need n of daily implied volatility                self.data = {}              # storing daily SKEW        self.contracts = {}         # storing option contracts        self.tickers_symbols = {}   # storing symbols under their tickers                self.symbol = self.AddEquity('SPY', Resolution.Daily).Symbol                self.day = -1        self.coarse_count = 50        self.selection_flag = False        self.UniverseSettings.Resolution = Resolution.Minute        self.AddUniverse(self.CoarseSelectionFunction, self.FineSelectionFunction)        self.SetSecurityInitializer(lambda x: x.SetDataNormalizationMode(DataNormalizationMode.Raw))        self.UniverseSettings.DataNormalizationMode = DataNormalizationMode.Raw                # rebalance on each Wednesday at 00:00 so it takes prices from Tuesday close        self.Schedule.On(self.DateRules.WeekStart(self.symbol, 1), self.TimeRules.BeforeMarketClose(self.symbol), self.Selection)
    def OnSecuritiesChanged(self, changes):        for security in changes.AddedSecurities:            security.SetFeeModel(CustomFeeModel(self))            security.SetLeverage(5)
    def OnData(self, data):        # each day store implied volatility for selected stocks        # if this is too slow, then add self.Time.minute == 31 to condition,        # because ~99% of subscribed options get implied volatility in this time each day                           # get option ticker from option symbol                                                         # based on option ticker get stock symbol                                                          # make sure, SKEW is updated once in a day                                                         # make sure, there are enough contracts for stock                                                          # get atm call and otm put contract                                                        # check if there are both contracts                                           # calculate stock's SKEW                                         # update RollingWindow for stock's SKEW values                                                  # execute once a day                       # check expiry of contracts                   # remove contract, when it is 10 days to expiry                                   # subscribe to contracts, if stock symbol doesn't have any                           # get new contracts after expiration                           # rebalance weekly                         # go through each stock in self.data and trade calculate mean of SKEW values                              # get SKEW values from RollingWindow and calculate their mean                                 # store stock's mean SKEW                           # make sure, there are enough stocks for quintile selection                    # sort stocks by mean SKEW value and perform quintile selection                   # long stocks with lowest mean SKEW               # go short stocks with highest mean SKEW                  # perform value weight calculation                      # trade execution         
            # data error handling