Overall Statistics |
Total Trades 0 Average Win 0% Average Loss 0% Compounding Annual Return 0% Drawdown 0% Expectancy 0 Net Profit 0% Sharpe Ratio 0 Probabilistic Sharpe Ratio 0% Loss Rate 0% Win Rate 0% Profit-Loss Ratio 0 Alpha 0 Beta 0 Annual Standard Deviation 0 Annual Variance 0 Information Ratio 0 Tracking Error 0 Treynor Ratio 0 Total Fees $0.00 |
""" Adaptive Volatility (position sizing) credit attribution: David Varadi https://cssanalytics.wordpress.com/2017/11/15/adaptive-volatility/ Aim: Get a better position sizing than [target_vol / realized_vol_{t-1}] (where the realized_vol is calculated over a fixed lookback period, e.g. past 20 days) using a more 'adaptive' volatility that varies its lookback period according to market conditions. The simplest method is to use the R-squared of the regression of prices vs time: 1. high R-squared indicates a trending market -> use short lookback periods to capture sudden changes in volatilities; 2. low R-squared instead iimplies a rangebound/mean-reverting market -> lengthen lookbacks since vol will revert to historical means. To translate the R_squared value into the alpha for an exponential moving average, the following exponential function is used (motivation: returns supposed lognormal): raw_alpha = exp[-10. * (1 - R_squared(price vs. time, period=20)] alpha = min(raw_alpha, 0.5) the 0.5 lower bound effectively limits the lookback to 3 days, since alpha := 2 / (1 + lookback). Such a capped aplha is used in an EMA of the squared returns for the past 20 days. Finally the (theoretical) daily exposure is: target_vol / sqrt( EMA_{t-1}(squared rturns, alpha) * 252) and target_vol is an annualised target vol, say 20%. To limit excessive trading, I only rebalace if theoretical exposure changes above a certain threshold (say 5%). Application hereby: long SPY (or similar) with a daily position sizing A more interesting use of this position sizing scheme is when using algorithms with long periodical rebalacings, say monthly or quarterly. """ import numpy as np import pandas as pd from datetime import datetime, timedelta from scipy.stats import linregress import decimal as d class AdaptiveVolatility(QCAlgorithm): def Initialize(self): self.SetStartDate(2019,1,1) self.SetCash(100000) #self.SetBrokerageModel(BrokerageName.InteractiveBrokersBrokerage, AccountType.Margin) symbols = [self.AddForex(ticker, Resolution.Minute).Symbol for ticker in ['EURUSD','USDCHF']] self.SetBenchmark('SPY') # schedule: rebalance dateRule = self.DateRules.EveryDay(symbols[0]) self.Schedule.On(dateRule, self.TimeRules.AfterMarketOpen(symbols[0], -90), self.rebalance) # schedule: email for recap at around close self.Schedule.On(dateRule, self.TimeRules.BeforeMarketClose(symbols[0], 0), self.JustBeforeMarketClose) self.back_period = 21 * 3 + 1 # 3 months self.vol_period = 21 # days for calc vol self.target_vol = 0.2 self.lev = 1.5 # max lev from ratio targ_vol / real_vol self.delta = 0.05 # min rebalancing self.w = 1. / len(symbols) self.x = np.asarray(range(self.vol_period)) ###################################### def rebalance(self): # get all weights try: pos_sizing = self.pos_sizing() except Exception as e: msg = f'Exception: {e}' self.Log(msg) return tot_port = self.Portfolio.TotalPortfolioValue for symbol, info in pos_sizing.items(): new_weight = info[0] yesterdayClose = info[1] security = self.Securities[symbol] quantity = security.Holdings.Quantity price = security.Price if price == 0: price = yesterdayClose # gauge if needs to trade (new weight vs. current one > self.delta) curr_weight = quantity * price / tot_port shall_trade = abs(new_weight - curr_weight) > self.delta if shall_trade: # self.SetHoldings(symbol, new_weight) delta_shares = int(new_weight * tot_port/ price) - quantity self.MarketOnOpenOrder(symbol, delta_shares) msg = f"{symbol} -- weight: {new_weight:.2f} (old weight was: {curr_weight:.2f}) -- last price: {price}" #self.Log(msg) def pos_sizing(self): # get daily returns for period = self.back_period allPrices = self.History(self.Securities.Keys, self.back_period, Resolution.Daily).close.unstack(level=0) pos = {} # calculate alpha for EWM for symbol in self.Securities.Keys: prices = allPrices[symbol] change = prices.pct_change().dropna() last = np.float(prices[-1]) rsq = self.rsquared(self.x, prices[-self.vol_period:]) alpha = min(0.5, np.exp(-10. * (1. - rsq))) vol = change.ewm(alpha=alpha).std() # alpha = 2/(span+1) = 1-exp(log(0.5)/halflife) ann_vol = np.float(vol.tail(1)) * np.sqrt(252) weight = (self.target_vol / ann_vol).clip(0.0, self.lev) * self.w # NB: self.w = 1/no_assets pos[symbol] = (weight, last) msg = f"{symbol}: {pos[symbol][0]}, rsqr: {rsq}, alpha: {alpha}, ann_vol = {ann_vol}" #self.Log(msg) return pos ###################################### def rsquared(self, x, y): # slope, intercept, r_value, p_value, std_err _, _, r_value, _, _ = linregress(x, y) return r_value**2 ###################################### ###################################### def OnMarginCallWarning(self): msg = f"{self.Time} : check warning margin call! Fast" self.Log(msg) ###################################### def JustBeforeMarketClose(self): msg = f"End of day: {self.Time} \nPortfolio value is {self.Portfolio.TotalPortfolioValue:.2f} and Margin Remaining is: {self.Portfolio.MarginRemaining:.2f} (Total Holdings Value: {self.Portfolio.TotalHoldingsValue:.2f})" self.Log(msg) ###################################### def OnOrderEvent(self, orderEvent): order = self.Transactions.GetOrderById(orderEvent.OrderId) self.Log(f"{self.Time}: {order.Type}: {orderEvent}") ###################################### def TimeIs(self, day, hour, minute): return self.Time.day == day and self.Time.hour == hour and self.Time.minute == minute