Binning-jackknife error analysis

The error estimation by means of the formula (4.21) is in most cases laborious. First, the calculation of the autocorrelation time is time-consuming and tedious. Second, to estimate the statistical error of nonlinear functions of the mean f(0) requires the consideration of error propagation. Fortunately, both problems can easily be solved approximately by the binning methods to be described now [80 82]. 

If the series of M correlated data points M) is cut into K pieces, each con- 

Technically, this relation is much more convenient than (4.18), because an estimate for the autocorrelation function is not needed. For a given time series with a total of as much 

This is only hue for uncorrelated data. For correlated data, the binning condition cannot 

M reduces the number of bins K and because of this the variance ct2b will also decrease, 

---