Exactly true !
actually, I was wrong..
Out of this 1.6 million pooled sample, 27,000 people went on to get infected, and most of them were unvaccinated (21,000)
This is the whole population, not just the control group. However, what this statement obscures and the paragraph attempts to explain is the distribution of who was infected when. Over the entire period, on average without any change in the infection rate (not likely true) one would expect 4500 infections per month in this 1.6 million population. The vaccine protected the vaccinated for a while so they were mostly not in the infected groups in the early months.. however, in the later months, this study is implying that they became an increasing percentage of the infected as time went on.. until they became the majority of those infected each month. (Thus the negative protection quotient vs. unvaccinated)