The source discusses different explanations of the “highest” rank: some focus on omitting the respect and particle together; examples may also be given both with and without an overt subject. Shorter phrasing does not therefore create a separate intrinsic grade on every occasion.
The central conclusion is that an unrestricted-seeming resemblance and direct predication of the model combine to produce especially strong emphasis. With only one effect the intensity is intermediate, and with both respect and particle expressed it is more explicit.
This preserves the criterion for ranking. Making the highest degree depend merely on omitting the subject or using only one word would be a superficial classification.