In this contribution, we present an alternative to the broadly used bootstrap method in nonlife insurance reserving based on state-space modeling of time series. The bootstrap method is useful when an insurance company aims to analyze the distribution of the reserves. However, the actual application of this method can be difficult. When the original portfolio of insurance claims is reasonably replicated, one can easily estimate the distribution of the reserve. Therefore, in this chapter, we apply the replication approach to the portfolio of insurance claims provided by Czech Insurers’ Bureau. This method supplements reserve estimation using state-space modeling of aggregated claims stacked to the form of a time series. In addition to that, we also use a method of clustering in order to split the data into several homogeneous groups. This should result in the estimation of more reliable reserves.