Summary class was often not updating the values for min_per_bin and using a
previous definition. This often resulted in subtracting the wrong per_bin
values when calculating delta-test-statistic and eneded up with negative values
for the test statistic. This fixes the problem by always redefining
min_per_bin.
Summary class was often not updating the values for
min_per_bin
and using a previous definition. This often resulted in subtracting the wrong per_bin values when calculating delta-test-statistic and eneded up with negative values for the test statistic. This fixes the problem by always redefiningmin_per_bin
.