Statistical and time domain analysis of ELMs in Joint European Torus (JET) plasmas is providing fresh insights into the ELMing process. These plasmas are often well adapted to such studies because of their long duration, such that large numbers of ELMs occur under quasi-stationary conditions. Understanding the ELMing process is a key challenge in magnetic confinement fusion plasma physics given the correlation between ELMing and enhanced confinement regimes, and the constraints on ELM magnitudes in future ITER plasmas. The ELMing process is a multiscale nonlinear plasma phenomenon and information on the underlying process can be acquired by analysing the distribution of events, here, ELM occurrences. This was noted initially in [1,2]. Subsequent advances in the scale and quality of the ELM data from JET, in particular, have led to rapid recent progress. Studies have involved: transitions in Type I ELMing in a sequence of similar JET plasmas in response to different gas puffing rates [3]; differentiating between Type I and Type III ELMs from first principles, in terms of extreme value distributions of inter-ELM time intervals [4]; and the identification of strong (and wholly unexpected) structure in the distribution of inter-ELM time intervals from a week-long sequence of quasi-identical JET plasmas [5]. Here we turn to a newly identified [6] correlation between ELM occurrence times in JET and the time evolution of a global measure of the plasma state. The latter is provided by the full flux loop VLD2 and VLD3 measurements, which are proportional to the changing poloidal magnetic flux through loops that encircle JET toroidally near the divertor coil system.