def cagr(qf_series: QFSeries, frequency=None): """ Returns the Compound Annual Growth Rate (CAGR) calculated for the given series. Parameters ---------- qf_series: QFSeries series of returns of an asset frequency: Frequency Frequency of the timeseries of returns; if it is None (by default) it is inferred from the timeseries of returns Returns ------- float annual compound return for the given series of prices """ prices_tms = qf_series.to_prices(frequency=frequency, initial_price=1.0) last_date = prices_tms.index[-1] first_date = prices_tms.index[0] period_length = last_date - first_date period_length_in_years = to_days(period_length) / DAYS_PER_YEAR_AVG total_return = prices_tms[-1] / prices_tms[0] - 1 return annualise_total_return( total_return=total_return, period_length_in_years=period_length_in_years, returns_type=SimpleReturnsSeries)
def avg_nr_of_trades_per1y(trades_returns: QFSeries, start_date: datetime, end_date: datetime): """ Calculates average number of trades per year for a given data-frame of trades. """ period_length = end_date - start_date period_length_in_years = to_days(period_length) / DAYS_PER_YEAR_AVG avg_number_of_trades_1y = len(trades_returns) / period_length_in_years return avg_number_of_trades_1y
def avg_nr_of_trades_per1y(trades: QFDataFrame, start_date: datetime, end_date: datetime): """ Calculates average number of trades per year for a given data-frame of trades. """ returns = trades[TradeField.Return] period_length = end_date - start_date period_length_in_years = to_days(period_length) / DAYS_PER_YEAR_AVG avg_number_of_trades_1y = returns.count() / period_length_in_years return avg_number_of_trades_1y
def _add_statistics_table(self): table = Table(column_names=["Measure", "Value"], css_class="table stats-table") number_of_trades = self.returns_of_trades.count() table.add_row(["Number of trades", number_of_trades]) period_length = self.end_date - self.start_date period_length_in_years = to_days(period_length) / DAYS_PER_YEAR_AVG avg_number_of_trades = number_of_trades / period_length_in_years / self.nr_of_assets_traded table.add_row(["Avg number of trades per year per asset", avg_number_of_trades]) positive_trades = self.returns_of_trades[self.returns_of_trades > 0] negative_trades = self.returns_of_trades[self.returns_of_trades < 0] percentage_of_positive = positive_trades.count() / number_of_trades percentage_of_negative = negative_trades.count() / number_of_trades table.add_row(["% of positive trades", percentage_of_positive * 100]) table.add_row(["% of negative trades", percentage_of_negative * 100]) avg_positive = positive_trades.mean() avg_negative = negative_trades.mean() table.add_row(["Avg positive trade [%]", avg_positive * 100]) table.add_row(["Avg negative trade [%]", avg_negative * 100]) best_return = max(self.returns_of_trades) worst_return = min(self.returns_of_trades) table.add_row(["Best trade [%]", best_return * 100]) table.add_row(["Worst trade [%]", worst_return * 100]) max_dd = max_drawdown(self.returns_of_trades) table.add_row(["Max drawdown [%]", max_dd * 100]) prices_tms = self.returns_of_trades.to_prices() total_return = prices_tms.iloc[-1] / prices_tms.iloc[0] - 1 table.add_row(["Total return [%]", total_return * 100]) annualised_ret = annualise_total_return(total_return, period_length_in_years, SimpleReturnsSeries) table.add_row(["Annualised return [%]", annualised_ret * 100]) avg_return = self.returns_of_trades.mean() table.add_row(["Avg return of trade [%]", avg_return * 100]) std_of_returns = self.returns_of_trades.std() table.add_row(["Std of return of trades [%]", std_of_returns * 100]) # System Quality Number sqn = avg_return / std_of_returns table.add_row(["SQN", sqn]) table.add_row(["SQN for 100 trades", sqn * 10]) # SQN * sqrt(100) table.add_row(["SQN * Sqrt(avg number of trades per year)", sqn * sqrt(avg_number_of_trades)]) self.document.add_element(table)
def _objective_function(self, trades: QFDataFrame): """ Calculates the simple SQN * sqrt(average number of trades per year) """ number_of_instruments_traded = len(self.all_tickers_tested) returns = trades[TradeField.Return] period_length = self.backtest_summary.end_date - self.backtest_summary.start_date period_length_in_years = to_days(period_length) / DAYS_PER_YEAR_AVG avg_number_of_trades_1y = returns.count() / period_length_in_years / number_of_instruments_traded sqn = returns.mean() / returns.std() sqn = sqn * np.sqrt(avg_number_of_trades_1y) return sqn
def test_to_days(self): expected = 3 actual = to_days(Timedelta('3 days')) self.assertEqual(expected, actual)