I read about a phalanx of maybe 16 ranks and I wonder if the rear ranks were really effective. Ardant du Picq in "Battle Studies" claimed that the rear ranks of a phalanx were ineffective as realistically only the two front ranks would be involved, while the rear would not push the front ranks forward but retreat if they were repulsed, thus not contributing to the fight.

Admittedly Du Picq is a very dated source, but his objections seem reasonable. Does modern scholarship disagree with this view, and if so why? Did any ancient sources comment on how many ranks a formation should have, or the effect of deep vs shallow ranks?