On WALS' chapter on verbal person marking alignment, it classifies English as having accusative alignment. Later on it discusses a 'restricted' alignment split between first and second person and third person, where first and second person have neutral alignment and third person accusative. I understand this, but not why English is then classed as having accusative alignment instead of neutral when 2/3 of the persons are neutral.
The quote from the chapter explains the mechanics of the classification. Split alignment with neutral and non-neutral alignments are classified under the non-neutral alignments. Classifying all split alignments as neutral would clearly not be very informative - although having details on the split might be useful.
But more importantly, you seem to be confusing accusative marking in English with agreement. Person is not the primary place of neutrality, rather it is the nominal vs. the pronominal system. I > me, we > us, he/she > him/her, they > them, who > whom. In such a system, you can't really talk of the second person being neutral but rather having the same accusative and nominative form - this zero form is common even in language with complex nominal accusative morphologies like Russian and nobody would call them anything but strictly accusative systems.
Because the accusative marking survived so prominently in the pronouns in English, it is pretty obvious that the syntax has primarily accusative alignment which is unmarked on nouns. There is other evidence for this in the passive form: The ball broke the window. > The window was broken by the ball. Where the roles are marked positionally and analytically.