I'm willing to bet I know more about this than you do, having done a couple of Institute of Statisticians diplomas and being a Bachelor of Science with Honours*, which required entry level Biology + another science (being Chemistry in my case, plus Economics thrown in) plus extensive probability theory and forecasting in my postgraduate business masters.
*This entailed a dissertation plus fifteen self-designed laboratory reports, plus
a Statistics paper in the finals.
The null hypothesis in the Estonia case ought to have been, 'We strongly suspect the bow visor falling off and seawater flooding the car deck, as in the Herald of Free Enterprise , so to test this hypothesis, we'll assume that it was NOT the bow visor falling off that was the prime cause of the accident. That is, we will investigate whether the massive hole in the starboard caused by what looks like an enormous force or the loud explosion type noises heard by 48% of the vanishingly small number of survivors around about the stroke of Swedish midnight, together with the reported blackout of radio and VHF signals, might have contributed. Only once we have eliminated these can we reject our null hypothesis that it was not the bow visor falling off because of a few strong waves.