Reckoning the Easter tables in 525, Dionysius Exiguus refused to go on numbering the years by Diocletian the persecutor and counted from the incarnation instead — and he began at 1, because Roman arithmetic had no zero to give a year. So 1 BC is followed straight by AD 1, and every span across the boundary is a year shorter than subtraction suggests. He also set the incarnation a few years late: Herod was alive at the birth in Matthew 2 and died in 4 BC, so the Nativity this dataset dates to c. 5 BC falls before the era named for it. Bede carried the reckoning into general use two centuries later, and the West has counted this way since.
On Easter, or the Paschal Cycle (Liber de Paschate) · Praefatio