Binning (or bunching) consists in grouping the terms of a statistical series in order to reduce its autocorrelation time. If the bin size is large enough, the samples of the binned series are uncorrelated; as a result, its error bar can be computed easily, i.e

\[\Delta \langle X \rangle = \sqrt{\frac{\tilde{\sigma}^2} {\tilde{N}-1}}\]

where \(\tilde{\sigma}^2\) and \(\tilde{N}\) are respectively the variance and size of the binned series.


make_binned_series(T series, int bin_size)
  • series: object with TimeSeries concept
  • bin_size: size of the bin

returns the binned time series.


#include <triqs/statistics.hpp>
using namespace triqs::statistics;
int main() {
  observable<double> A;
  A << 1.;
  A << 1.5;
  A << .2;
  A << 1.1;
  auto A_b = make_binned_series(A, 2);
  std::cout << A_b << std::endl;
  return 0;