Recent progress in forensic genetics has introduced a number of closely located short tandem repeat (STR) markers on the X chromosome. Inevitably, dependencies arise that have to be accounted for. This paper will in detail explore the complex statistical interpretation of X-chromosomal STR markers, focusing on likelihood calculations. Specifically, we will investigate how the phase uncertainty of haplotypes comes into play in the statistical evaluations and what curious effects this phenomenon can have. The starting point is the different real cases where the weight of evidence has provided unexpected results that require further investigation in order to be fully understood. We will touch upon subjects such as association between alleles, recombinations, and mutations. The aim of this study is to facilitate a better understanding of the interaction between the concepts in addition to provide an understanding why good estimates of haplotype frequencies are crucial. The individual subjects have been discussed in other fields, whereas this study will focus on forensic applications where few studies have been conducted relating to the understanding of how these concepts interact.