Here's a bit of an "out there" hypothesis. Imagine you had a bit of grime in the spigot that had been happily growing bacteria since the last time the bottling bucket was used. Now imagine liquid starts flowing through there. The first few bottles would dissolve a lot of the gunk, carrying the bacteria with it into the bottles. As the process proceeded, because the gunk is soluble, there'd be less until eventually there was none. So those first bottles would be infected, the last would not. Admittedly, it's a wild-a**ed guess but plausible... There's always bacteria in beer but as with Covid, it's the load that counts.