Merge pull request #859 from quantopian/future_beta

Better handle missing benchmarks when calculating beta

docs/source/whatsnew/0.8.4.txt

@@ -44,6 +44,9 @@ Bug Fixes
   the ``len`` of a :class:`~zipline.protocol.SIDData` object. This would cause
   us to think that the object was not empty even when it was (:issue:`826`).
 
+* Fixes an error raised in calculating beta when benchmark data were sparse.
+  Instead `numpy.nan` is returned (:issue:`859`).
+
 Performance
 ~~~~~~~~~~~
 

tests/risk/test_risk_period.py

@@ -18,6 +18,10 @@ import datetime
 import calendar
 import numpy as np
 import pytz
+
+from itertools import chain
+from six import itervalues
+
 import zipline.finance.risk as risk
 from zipline.utils import factory
 
@@ -622,3 +626,17 @@ class TestRisk(unittest.TestCase):
             )
             self.assert_month(start_date.month, col[-1].end_date.month)
             self.assert_last_day(col[-1].end_date)
+
+    def test_sparse_benchmark(self):
+        benchmark_returns = self.benchmark_returns_06.copy()
+        # Set every other day to nan.
+        benchmark_returns.iloc[::2] = np.nan
+
+        report = risk.RiskReport(
+            self.algo_returns_06,
+            self.sim_params,
+            benchmark_returns=benchmark_returns,
+            env=self.env,
+        )
+        for risk_period in chain.from_iterable(itervalues(report.to_dict())):
+            self.assertIsNone(risk_period['beta'])

zipline/finance/risk/period.py

@@ -252,13 +252,19 @@ class RiskMetricsPeriod(object):
         http://en.wikipedia.org/wiki/Beta_(finance)
         """
         # it doesn't make much sense to calculate beta for less than two days,
-        # so return none.
+        # so return nan.
         if len(self.algorithm_returns) < 2:
-            return 0.0, 0.0, 0.0, 0.0, []
+            return np.nan, np.nan, np.nan, np.nan, []
 
         returns_matrix = np.vstack([self.algorithm_returns,
                                     self.benchmark_returns])
         C = np.cov(returns_matrix, ddof=1)
+
+        # If there are missing benchmark values, then we can't calculate the
+        # beta.
+        if not np.isfinite(C).all():
+            return np.nan, np.nan, np.nan, np.nan, []
+
         eigen_values = la.eigvals(C)
         condition_number = max(eigen_values) / min(eigen_values)
         algorithm_covariance = C[0][1]