FIX refactor annual_return to return CAGR. Removes returns_style para…

pyfolio/plotting.py

@@ -490,12 +490,12 @@ def show_perf_stats(returns, factor_returns, live_start_date=None):
         returns_live = returns[returns.index > live_start_date]
 
         perf_stats_live = np.round(timeseries.perf_stats(
-            returns_live, returns_style='arithmetic',
+            returns_live,
             factor_returns=factor_returns), 2)
         perf_stats_live.columns = ['Out_of_Sample']
 
         perf_stats_all = np.round(timeseries.perf_stats(
-            returns, returns_style='arithmetic',
+            returns,
             factor_returns=factor_returns), 2)
         perf_stats_all.columns = ['All_History']
 
@@ -508,7 +508,7 @@ def show_perf_stats(returns, factor_returns, live_start_date=None):
           str(int(len(returns_backtest) / APPROX_BDAYS_PER_MONTH)))
 
     perf_stats = np.round(timeseries.perf_stats(
-        returns_backtest, returns_style='arithmetic',
+        returns_backtest,
         factor_returns=factor_returns), 2)
     perf_stats.columns = ['Backtest']
 
@@ -1190,7 +1190,7 @@ def plot_slippage_sensitivity(returns, transactions, positions,
     for bps in range(1, 100):
         adj_returns = txn.adjust_returns_for_slippage(returns, turnover, bps)
         avg_returns = timeseries.annual_return(
-            adj_returns, style='calendar')
+            adj_returns)
         avg_returns_given_slippage.loc[bps] = avg_returns
 
     avg_returns_given_slippage.plot(alpha=1.0, lw=2, ax=ax)

pyfolio/tests/test_timeseries.py

@@ -40,8 +40,8 @@ def test_get_max_drawdown(
         # Need to use isnull because the result can be NaN, NaT, etc.
         self.assertTrue(
             pd.isnull(peak)) if expected_peak is None else self.assertEqual(
-            peak,
-            expected_peak)
+                peak,
+                expected_peak)
         self.assertTrue(
             pd.isnull(valley)) if expected_valley is None else \
             self.assertEqual(
@@ -235,21 +235,20 @@ class TestStats(TestCase):
         [10, -10, 10]) / 100.  # Ends in drawdown
     dt = pd.date_range('2000-1-3', periods=3, freq='D')
 
+    px_list_2 = [1.0, 1.2, 1.0, 0.8, 0.7, 0.8, 0.8, 0.8]
+    dt_2 = pd.date_range('2000-1-3', periods=8, freq='D')
+
     @parameterized.expand([
-        (simple_rets, 'calendar', utils.DAILY, 0.10584000000000014),
-        (simple_rets, 'compound', utils.DAILY, 0.16317653888658334),
-        (simple_rets, 'calendar', utils.DAILY, 0.10584000000000014),
-        (simple_rets, 'compound', utils.DAILY, 0.16317653888658334),
-        (simple_week_rets, 'compound', utils.WEEKLY, 0.031682168889005213),
-        (simple_week_rets, 'calendar', utils.WEEKLY, 0.021840000000000033),
-        (simple_month_rets, 'compound', utils.MONTHLY, 0.0072238075842128158),
-        (simple_month_rets, 'calendar', utils.MONTHLY, 0.0050400000000000071)
+        (simple_rets, utils.DAILY, 0.15500998835658075),
+        (simple_week_rets, utils.WEEKLY, 0.030183329386562319),
+        (simple_month_rets, utils.MONTHLY, 0.006885932704891129)
     ])
-    def test_annual_ret(self, returns, style, period, expected):
+    def test_annual_ret(self, returns, period, expected):
         self.assertEqual(
             timeseries.annual_return(
                 returns,
-                style=style, period=period),
+                period=period
+            ),
             expected)
 
     @parameterized.expand([
@@ -306,16 +305,13 @@ def test_beta(self, returns, benchmark_rets, rolling_window, expected):
             expected)
 
     @parameterized.expand([
-        (pd.Series(px_list,
-                   index=dt), 'calendar', -8.3999999999999559),
-        (pd.Series(px_list,
-                   index=dt), 'arithmetic', 84.000000000000014)
+        (pd.Series(px_list_2,
+                   index=dt_2).pct_change().dropna(), -2.3992211554712197)
     ])
-    def test_calmar(self, returns, returns_style, expected):
+    def test_calmar(self, returns, expected):
         self.assertEqual(
             timeseries.calmar_ratio(
-                returns,
-                returns_style=returns_style),
+                returns),
             expected)
 
     @parameterized.expand([

pyfolio/timeseries.py

@@ -191,20 +191,14 @@ def max_drawdown(returns):
     return -1 * MDD
 
 
-def annual_return(returns, style='compound', period=DAILY):
+def annual_return(returns, period=DAILY):
     """Determines the annual returns of a strategy.
 
     Parameters
     ----------
     returns : pd.Series
         Periodic returns of the strategy, noncumulative.
          - See full explanation in tears.create_full_tear_sheet.
-    style : str, optional
-        - If 'compound', then return will be calculated in geometric
-          terms: (1+mean(all_daily_returns))^252 - 1.
-        - If 'calendar', then return will be calculated as
-          ((last_value - start_value)/start_value)/num_of_years.
-        - Otherwise, return is simply mean(all_daily_returns)*252.
     period : str, optional
         - defines the periodicity of the 'returns' data for purposes of
         annualizing. Can be 'monthly', 'weekly', or 'daily'
@@ -213,7 +207,7 @@ def annual_return(returns, style='compound', period=DAILY):
     Returns
     -------
     float
-        Annual returns.
+        Annual Return as CAGR (Compounded Annual Growth Rate)
 
     """
 
@@ -230,16 +224,15 @@ def annual_return(returns, style='compound', period=DAILY):
             )
         )
 
-    if style == 'calendar':
-        num_years = len(returns) / ann_factor
-        df_cum_rets = cum_returns(returns, starting_value=100)
-        start_value = df_cum_rets[0]
-        end_value = df_cum_rets[-1]
-        return ((end_value - start_value) / start_value) / num_years
-    if style == 'compound':
-        return pow((1 + returns.mean()), ann_factor) - 1
-    else:
-        return returns.mean() * ann_factor
+    num_years = float(len(returns)) / ann_factor
+    df_cum_rets = cum_returns(returns, starting_value=100)
+    start_value = 100
+    end_value = df_cum_rets[-1]
+
+    total_return = (end_value - start_value) / start_value
+    annual_return = (1. + total_return) ** (1 / num_years) - 1
+
+    return annual_return
 
 
 def annual_volatility(returns, period=DAILY):
@@ -278,7 +271,7 @@ def annual_volatility(returns, period=DAILY):
     return returns.std() * np.sqrt(ann_factor)
 
 
-def calmar_ratio(returns, returns_style='calendar', period=DAILY):
+def calmar_ratio(returns, period=DAILY):
     """
     Determines the Calmar ratio, or drawdown ratio, of a strategy.
 
@@ -287,8 +280,6 @@ def calmar_ratio(returns, returns_style='calendar', period=DAILY):
     returns : pd.Series
         Daily returns of the strategy, noncumulative.
          - See full explanation in tears.create_full_tear_sheet.
-    returns_style : str, optional
-        See annual_returns' style
     period : str, optional
         - defines the periodicity of the 'returns' data for purposes of
         annualizing. Can be 'monthly', 'weekly', or 'daily'
@@ -309,7 +300,6 @@ def calmar_ratio(returns, returns_style='calendar', period=DAILY):
     if temp_max_dd < 0:
         temp = annual_return(
             returns=returns,
-            style=returns_style,
             period=period
         ) / abs(max_drawdown(returns=returns))
     else:
@@ -368,8 +358,6 @@ def sortino_ratio(returns, required_return=0, period=DAILY):
     returns : pd.Series or pd.DataFrame
         Daily returns of the strategy, noncumulative.
          - See full explanation in tears.create_full_tear_sheet.
-    returns_style : str, optional
-        See annual_returns' style
     required_return: float / series
         minimum acceptable return
     period : str, optional
@@ -459,8 +447,6 @@ def sharpe_ratio(returns, risk_free=0, period=DAILY):
     returns : pd.Series
         Daily returns of the strategy, noncumulative.
          - See full explanation in tears.create_full_tear_sheet.
-    returns_style : str, optional
-        See annual_returns' style
     period : str, optional
         - defines the periodicity of the 'returns' data for purposes of
         annualizing. Can be 'monthly', 'weekly', or 'daily'
@@ -728,7 +714,6 @@ def calc_alpha_beta(returns, factor_returns):
 
 def perf_stats(
         returns,
-        returns_style='compound',
         return_as_dict=False,
         factor_returns=None,
         period=DAILY):
@@ -740,8 +725,6 @@ def perf_stats(
     returns : pd.Series
         Daily returns of the strategy, noncumulative.
          - See full explanation in tears.create_full_tear_sheet.
-    returns_style : str, optional
-       See annual_returns' style
     return_as_dict : boolean, optional
        If True, returns the computed metrics in a dictionary.
     period : str, optional
@@ -761,15 +744,11 @@ def perf_stats(
     """
 
     all_stats = OrderedDict()
-    all_stats['annual_return'] = annual_return(
-        returns,
-        style=returns_style, period=period)
+    all_stats['annual_return'] = annual_return(returns, period=period)
     all_stats['annual_volatility'] = annual_volatility(returns, period=period)
     all_stats['sharpe_ratio'] = sharpe_ratio(
         returns)
-    all_stats['calmar_ratio'] = calmar_ratio(
-        returns,
-        returns_style=returns_style, period=period)
+    all_stats['calmar_ratio'] = calmar_ratio(returns, period=period)
     all_stats['stability'] = stability_of_timeseries(returns)
     all_stats['max_drawdown'] = max_drawdown(returns)
     all_stats['omega_ratio'] = omega_ratio(returns)