The table that does it that way is a complete mess but I guess it was a good idea and got some things right.

They first started doing it with valence shells in 1864, Mendeleev had a pretty close one based on atomic weights in 1871 and correctly predicted that there were missing elements based on valences + weights. The atomic numbers which determine the valences appear to have been discovered about 40 years later.

I guess you are correct but I think the question was about the modern table.