Buy article online - an online subscription or single-article purchase is required to access this article.
![j logo](https://journals.iucr.org/logos/jicons/j_36x36.png)
The statistical properties of intensities affected by counting loss based on conventional non-extended and extended dead-time models are examined by a Monte Carlo method. It has been confirmed that the variance of the counted pulses for the non-extended dead-time model with the rate of generated pulses r and the dead-time τ is given by
, while that for the extended dead-time model is given by
, as proposed by Laundy & Collins [(2003). J. Synchrotron Rad. 10, 214–218], for the mean values of counted pulses μnon and μext, respectively. Practical formulae to estimate the statistical errors of the corrected intensities are also presented.
![\sigma_{\rm non}^2 = \mu_{\rm non}/(1+r \tau)^2](http://journals.iucr.org/j/issues/2007/05/00/fe5030//teximages/fe5030fi1.gif)
![\sigma_{\rm ext}^2 = \mu_{\rm ext} [1 - 2r\tau \exp(-r \tau)]](http://journals.iucr.org/j/issues/2007/05/00/fe5030//teximages/fe5030fi2.gif)