About this model: Segment count is modeled as Poisson-distributed and segment lengths as exponentially distributed, calibrated per relationship category from published average total-cM figures. The largest segment is evaluated using the distribution of the maximum of N exponential variables (order statistics) rather than an arbitrary bell curve. This is a genuine improvement over simple cM lookup tables, but it is still a simplified model: it does not separately account for IBD2 sharing (e.g. in full siblings), multiple shared ancestral lines (endogamy or pedigree collapse), or the sex-specific recombination differences on the X chromosome. Relationship categories with the same expected relatedness (e.g. half-sibling, grandparent/grandchild, and aunt/uncle) will always overlap heavily — no amount of segment data fully separates them. Always confirm a specific relationship using family trees, ages, shared matches, and documentary research alongside DNA evidence.