When using accumulators its extra work to always catch and ignore errors because the variance accumulator needs at least two elements, the QuantileEstimator needs 5 (but that is variable). So it maybe beneficial, if accumulator.value does not throw an exception, but instead returns float('nan') if the value cannot be calculated yet. This is also automatically ignored by matplotlib during the plot process.
When using accumulators its extra work to always catch and ignore errors because the variance accumulator needs at least two elements, the
QuantileEstimatorneeds 5 (but that is variable). So it maybe beneficial, ifaccumulator.valuedoes not throw an exception, but instead returnsfloat('nan')if the value cannot be calculated yet. This is also automatically ignored by matplotlib during the plot process.