But consider the situation where the EV is 15 stops lower. Instead of capturing 2^15 and 2^16 photons per pixel respectively, the two sensors now capture 1 and 2 photons per pixel respectively. To my mind that says that the small-pixel sensor has reached its limit - one photon is the minimum which cave counted and it cannot record faiinter scenes. However the large-pixel sensor can discriminate between 2 pixels and 1 pixel, and it can keep going down to one more stop of EV. In other words, it has a greater dynamic range.