Palaeointenisty measurements are vital for understanding the deep Earth and the evolution of the geomagnetic field, however, they suffer from non-ideal behaviour that compromises the fidelity of the results. In an attempt to screen for the non-ideal behaviour that plagues palaeointensity experiments, selection criteria are used. One such non-ideal behaviour is partial thermoremanent magnetisation (pTRM) tails, which in the IZZI protocol result in a zig-zagging pattern in the Arai plot. While pTRM tail checks can be carried out, these are not routinely used in IZZI experiments. We set out a new framework with which to create, develop and re-assess selection criteria before applying it to our proposed new selection statistic, Ziggie, which is specific to detecting zig-zagging behaviour. Ziggie meets all the requirements set out; that is, it is able to objectively and robustly identify the target behaviour, consistently provide a quantifiable assessment of the target behaviour, and is predominately sensitive to the target behaviour while being free of undue influences. Other pre-existing statistics for zig-zagging in Arai plots, Z*, IZZI_MD, β and SCAT, are also tested, but fail to meet all requirements. We find that Ziggie should be adopted in selection criteria sets for the IZZI protocol with a criterion of Ziggie ≤ 0.1.